CHAPTER VAX PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX Architecture Subset Rules . . . . . . . . . VAX-3 POLYD and POLYF Errors on 78x Processors . . . . VAX-7 DECUS European VAX SIG Question and Answer Session . . . . . . . . . . . . . . . . . . . VAX-17 Fall 1987 US VAX Systems SIG SIR Ballot Results VAX-74 INPUT/OUTPUT . . . . . . . . . . . . . . . . . VAX-98 VAX System SIG Committee List . . . . . . . . VAX-133 Forms at the End INPUT/OUTPUT Submission Form . . . . . . . . . VAX-138 System Improvement Request Submission Form . . VAX-140 PAGESWAPPER - January 1988 - Volume 9 Number 6 To register for on-line submission to the Pageswapper dial: (617) 262-6830 (in the United States) using a 1200 baud modem and log in with the username PAGESWAPPER. Articles for publication in the Pageswapper can be sent (US mail only -- no "express" services please) to: Larry Kilgallen, PAGESWAPPER Editor Box 81, MIT Station Cambridge, MA 02139-0901 USA Preference is given to material submitted as machine-readable text (best is Runoff source). Line length should not exceed 64 characters and the number of text lines per page should not exceed 48 (these limits are particularly important for sample commands, etc. where simple text justification will not produce a meaningful result). Please do not submit program source, as that is better distributed on the VAX SIG tape. Please do not submit "slides" from DECUS Symposia presentations (or other meetings) as they are generally a very incomplete treatment for those readers of the Pageswapper who are not so fortunate as to be able to travel to Symposia. Please DO write articles based on such slides to get the content across to a wider audience than is able to attend. Change of address, reports of non-receipt, and other circulation correspondence should be sent to: DECUS U.S. Chapter Attention: Publications Department 249 Northboro Road (BPO2) Marlborough, MA 01752 USA Only if discrepancies of the mailing system are reported can they be analyzed and corrected. VAX-2 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX Architecture Subset Rules VAX Architecture Subset Rules VAX ARCHITECTURE MANAGEMENT DIGITAL EQUIPMENT CORPORATION BXB1-1/E11 85 SWANSON ROAD BOXBOROUGH, MA 01719-1326 Abstract: VAX computers have the option either to implement all of the instructions in the VAX instruction set in hardware or microcode, or to implement only a subset of the instructions in hardware or microcode and emulate the remaining instructions with operating-system software. VAX computers announced after 1986 will use a new set of rules to determine which instructions are to be implemented in hardware or microcode, and which may be emulated by operating-system level software supplied by Digital Equipment Corporation. Introduction and History: In order to design and manufacture a VAX computer, it is necessary to make trade-offs between marketing and development deadlines, development costs (both hardware and software), manufacturing costs, and system performance. Sometimes these trade-offs suggest that the processor should implement only a subset of the VAX architecture. For example, in order to make feasible the development and manufacturing of the MicroVAX I and MicroVAX II, only a subset of the VAX instruction set was implemented in hardware and microcode. The remaining instructions were emulated by operating-system level software. Although this process reduced the performance of the instructions that were emulated, it allowed the MicroVAX I and MicroVAX II to be designed and manufactured in a timely manner. The use of emulation software enabled these systems to present a complete architecture. New Instruction Subsets: Due to the specific trade-offs associated with the MicroVAX I and MicroVAX II, each of these systems implements and emulates a different group of instructions. In an effort to determine a standard set of instructions that may be emulated on future VAX computers, Digital Equipment Corporation began a study in 1986. The results of this study recommended that instructions which fit the following criteria be allowed to be emulated on future VAX computers. VAX-3 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX Architecture Subset Rules o The instruction must not be extensively used by system software (VMS, Run Time Library, ULTRIX, ELN, etc.) or extensively generated by Digital's compilers. o The instruction must be difficult to implement, either requiring a great deal of extra hardware or microcode. The instructions that are allowed to be emulated on future VAX computers are listed later in this article. Effect on User Software: Since there will still be support for every instruction (whether implemented in hardware/microcode or emulated) on every VAX computer, the new instruction subset rules have a minimal effect on user-level software. It is anticipated that low-end VAX computers will emulate most of the allowed instructions. However, mid-range and high-end VAX computers will continue to implement in hardware/microcode a majority or all of the allowed instructions. Those who write programs in high-level languages (such as FORTRAN, COBOL, Pascal, Ada, etc.) will see little if any change, and should not alter anything in their programs. Nevertheless, it is possible that those who write time-critical applications in MACRO (VAX Assembly Language) on low-end VAX computers may experience a slight change in the performance of the application, if the instructions that they are using are being emulated. Instructions allowed to be emulated: VAX computers announced after 1986 are required to follow the guidelines below to determine which instructions may be emulated. Emulation software will be included in the system-level software supplied by Digital Equipment Corporation. The VAX Architecture Reference Manual (Digital Press, Order Number EY-3459E-DP) lists those VAX computers which utilize instruction emulation, along with the instructions that are emulated. 1. Instructions in an application extension group may be implemented or omitted only as a group. That is, if any instruction in one of the application groups is implemented (omitted), all instructions in that application extension group are implemented (omitted). All VAX computers announced after 1986 have the option to implement zero or more of the application extension groups. Digital Equipment Corporation will supply emulation software for any instructions that are in omitted application groups. VAX compilers will support features to complement the application extension groups VAX-4 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX Architecture Subset Rules implemented in VAX computers. The application extension groups are: o A packed-decimal-string group with sixteen instructions: ADDP4 - Add Packed 4 Operand ADDP6 - Add Packed 6 Operand ASHP - Arithmetic Shift and Round Packed CMPP3 - Compare Packed 3 Operand CMPP4 - Compare Packed 4 Operand CVTLP - Convert Longword to Packed CVTPL - Convert Packed to Longword CVTPS - Convert Packed to Leading Separate Numeric CVTPT - Convert Packed to Trailing Numeric CVTSP - Convert Leading Separate Numeric to Packed CVTTP - Convert Trailing Numeric to Packed DIVP - Divide Packed MOVP - Move Packed MULP - Multiply Packed SUBP4 - Subtract Packed 4 Operand SUBP6 - Subtract Packed 6 Operand o An extended-accuracy group with twenty-nine H-floating and octaword instructions: ADDH2 - Add H_floating 2 Operand ADDH3 - Add H_floating 3 Operand CLRH - Clear H_floating (Octaword) (CLRO) CMPH - Compare H_floating CVTBH - Convert Byte to H_floating CVTWH - Convert Word to H_floating CVTLH - Convert Longword to H_floating CVTHB - Convert H_floating to Byte CVTHW - Convert H_floating to Word CVTHL - Convert H_floating to Longword CVTRHL - Convert Rounded H_floating to Longword CVTFH - Convert F_Floating to H_floating CVTDH - Convert D_floating to H_floating CVTGH - Convert G_floating to H_floating CVTHF - Convert H_floating to F_floating CVTHD - Convert H_floating to D_floating CVTHG - Convert H_floating to G_floating DIVH2 - Divide H_floating 2 Operand DIVH3 - Divide H_floating 3 Operand VAX-5 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX Architecture Subset Rules MNEGH - Move Negated H_floating MOVAH - Move Address H_floating (MOVAO) MOVH - Move H_floating MOVO - Move Octaword MULH2 - Multiply H_floating 2 operand MULH3 - Multiply H_floating 3 operand PUSHAH - Push Address H_floating (Octaword) (PUSHAO) SUBH2 - Subtract H_floating 2 Operand SUBH3 - Subtract H_floating 3 Operand TSTH - Test H_floating 2. No VAX computer announced after 1986 is required to implement instructions in the emulated-only instruction group. A VAX computer may implement (omit) any number of these instructions (that is, the computer may implement CRC and MATCHC only, and omit all the others). Digital Equipment Corporation will supply emulation software for any instructions that are omitted. Instructions in the emulated-only instruction group are: CRC - Cyclic Redundancy Check EDITPC - Edit Packed to Character String MATCHC - Match Characters MOVTC - Move Translated Characters MOVTUC - Move Translated Until Character ACBF - Add Compare and Branch F_floating ACBD - Add Compare and Branch D_floating ACBG - Add Compare and Branch G_floating ACBH - Add Compare and Branch H_floating EMODF - Extended Multiply and Integerize F_floating EMODD - Extended Multiply and Integerize D_floating EMODG - Extended Multiply and Integerize G_floating EMODH - Extended Multiply and Integerize H_floating POLYF - Polynomial Evaluation F_floating POLYD - Polynomial Evaluation D_floating POLYG - Polynomial Evaluation G_floating POLYH - Polynomial Evaluation H_floating VAX-6 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors POLYD and POLYF Errors on 78x Processors VAX ARCHITECTURE MANAGEMENT DIGITAL EQUIPMENT CORPORATION BXB1-1/E11 85 SWANSON ROAD BOXBOROUGH, MA 01719-1326 Abstract: Under certain obscure data conditions, execution of a POLYF or POLYD instruction on a VAX-11/780 or VAX-11/785 with the Floating Point Accelerator (FPA) installed will produce an answer that is different from the answer defined by the VAX architecture. These deviations from the VAX architecture are minor, and should not affect your programs. Software supplied by Digital Equipment Corporation does not generate any of these problems. There are no plans to fix these cases in future microcode or hardware releases at this time. This article describes the data conditions that result in answers that are different from those defined by the VAX Architecture. Introduction: The VAX architecture is a standard that all VAX computers (MicroVAX II, VAX 8800, VAX-11/780, etc.) follow. This standard allows correctly-written software to be executed on the entire line of VAX computers. The VAX-11/780 and VAX-11/785 deviate slightly from this standard, in respect to the POLYF and POLYD instructions. POLYF and POLYD are complex instructions that evaluate a polynomial given the degree, coefficients, and the argument. Under certain obscure conditions a POLYF or POLYD instruction, when executed on either a VAX-11/780 or VAX-11/785 with the FPA installed, will generate an answer that is different from that defined by the VAX architecture. These differences are minor in that they range from producing an answer that is slightly greater than the architecturally defined answer, to handling a floating underflow exception in a way different from that defined by the architecture. Because of the obscure data conditions that cause these differences, you should not be affected by them. They are being listed here only so that you may be aware of them. None of these cases have been reported by customers. They were all found by the exhaustive testing procedures of Digital Equipment Corporation. VAX-7 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors Background: Some knowledge about the format of floating point numbers and the internals of the POLY instruction is helpful in order to understand the data conditions that cause the errors. Floating Point Numbers: The VAX architecture includes four different types of floating point numbers, F_Floating, D_Floating, G_Floating, and H_Floating. Since the POLY deviations include only the F_Floating and D_Floating formats, descriptions of G_Floating and H_Floating will not be included. A floating point number is defined by three attributes, a sign, a mantissa, and an exponent. For example, let's consider the value -0.1875 (-3/16), in binary this can be represented as: 0 -0.0011 * 2 This is an example of an unnormalized representation. However, the VAX architecture requires the mantissa to be normalized, so that its most significant bit must be 1. This is accomplished by shifting the mantissa to the left and subtracting one from the exponent for every place that is shifted. So the normalized representation would be: -2 -0.11 * 2 For this example, the sign is negative (-1), the mantissa is .11, and the exponent is -2. The F_Floating and D_Floating formats use 32 and 64 bits, respectively, to store a corresponding floating-point value. The following table describes how many bits the F_Floating and D_Floating formats use to store each field. F_Floating D_Floating ----------------------- Sign 1 | 1 Mantissa 23 | 55 Exponent 8 | 8 ----------------------- Total 32 64 If two n-bit floating-point mantissas are added, the sum will be a maximum of n+1 bits. If two n-bit floating-point mantissas are multiplied, the product will be a maximum of 2*n bits. All of these bits are not used in the actual calculations internal VAX-8 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors to the computer. Only a group of high-order bits are retained, these bits are sufficient to maintain the calculation without having to store all of the bits generated. The bits that are retained are the high-order mantissa bits plus a few bits of lower significance which are labeled "guard bits." The POLY Instructions: The POLY instruction uses Horner's method to evaluate a polynomial, given the degree, the argument, and a pointer to a table of coefficients. For more information on the POLY instructions, refer to the "VAX Architecture Reference Manual" (Digital Press, Order number EY-3459E-DP). Assume that: o The Polynomial is of degree n. o Coefficients are of the form (from highest degree to lowest degree): C[n], C[n-1] . . . . C[0] o The Argument is x. o The Polynomial is of the form: n n-1 C[n]*x + C[n-1]*x + . . . . C[1]*x + C[0] The VAX-11/780 and VAX-11/785 follow these steps to implement Horner's method: Partial_Result = C[n] WHILE n > 0 DO !Multiplication Step Partial_Result = Partial_Result * Argument !After the Partial_Result has been multiplied by the !Argument, the result is truncated to 31 (POLYF) or !63 (POLYD) bits. This is referred to as the first !truncation step. !Add Step Partial_Result = Partial_Result + C[n-1] VAX-9 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors !After C[n-1] has been added to the Partial_Result, !the result is again truncated to 31 (POLYF) or 63 !(POLYD) bits. This is referred to as the second !truncation step. !The Partial_Result is now normalized. This is !referred to as the normalization step. !The Partial_Result is now rounded. This is referred !to as the rounding step. !Perform next iteration n = n -1 ENDDO !Store the result Answer = Partial_Result Conditions that Cause the Deviations: Currently there exist four situations when execution of a POLYF or a POLYD instruction on either a VAX-11/780 or a VAX-11/785 with the FPA installed will generate an answer that is different from the one produced by other VAX computers and defined by the VAX architecture. Each of these deviations occurs only when the FPA is installed on the VAX-11/780 and VAX-11/785. CASE 1 The first case concerns POLYF and POLYD reporting a floating underflow exception in the wrong manner. A floating underflow exception occurs if the instruction generated a result which consists of an exponent that is less than the smallest representable exponent (the smallest representable exponent is -127 for F_ and D_Floating numbers). If a floating underflow exception occurs, the VAX architecture defines that a zero is written as the result. Given the following data conditions, the VAX-11/780 and VAX-11/785 handle a floating underflow exception in a different manner. VAX-10 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors o If the lowest order coefficient (C0) is zero o and if underflow occurs during the multiplication before C0 is added o and if the product generated by this multiplication is negative, then the instruction will store a reserved operand, instead of storing a zero. A reserved operand is an operand that has a format that is reserved for future use by Digital Equipment Corporation. If a following instruction were to read this reserved operand (assuming it were valid data), another exception would occur, a reserved-operand exception. CASE 2 The second case concerns a rounding error in the execution of POLYD. If the following data conditions are met, the result of the operation will be rounded when it should not be. o If the result after the multiplication step is negative, o and if the least-significant bit of the mantissa is zero, o and if the most-significant guard bit is one, then the result generated will have a one in the least significant bit of the mantissa when it should be a zero. Thus, the magnitude of the mantissa is slightly greater than the architecturally defined result. CASE 3 The third case concerns a rounding error in POLYF and POLYD. The multiplication step should retain an unnormalized result of 31 bits for POLYF (24 high-order bits and 7 low-order guard bits) and 63 bits for POLYD (56 high-order bits and 7 low-order guard bits). If the following data conditions are met, the result of the operation will be rounded when it should not be. o If the guard bits equal exactly 1/4 the mantissa's least VAX-11 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors significant bit (for example, the least significant bit of the mantissa is a 1 and the guard bits are 0100000), o and the next coefficient to add is negative, then the result generated will have a one in the least significant bit of the mantissa when it should be a zero. Thus, the magnitude of the mantissa is slightly greater than the architecturally defined result. CASE 4 The fourth case involves a truncation problem in the POLYF instruction. The VAX architecture allows a VAX computer to retain either 31 or 32 bits of precision after a truncation step. However, the VAX computer must retain the same number of bits (either 31 or 32) all of the time; it cannot retain 32 bits sometimes and 31 bits the other times. The VAX-11/780 and VAX-11/785 were designed to retain 31 bits. If the following data conditions are met, the VAX-11/780 and VAX-11/785 will retain 32 bits of precision instead of 31. o If the sign of the next coefficient to add is opposite to that of the current partial result, o and if bits <30:28> of the sum are all the same (000 or 111), o and if bit 31 of the sum is one, then the VAX-11/780 and VAX-11/785 will retain 32 bits of precision instead of 31. Examples that Reflect the Data Conditions: CASE 1 Example: ; The POLYD instruction below includes an underflow condition ; that occurs during the last multiplication step. The result ; would be negative if not for the underflow at the last ; iteration. The last coefficient (C0) is zero. Under these ; conditions, the result is left containing a reserved operand, ; and, even if PSL is set, a floating underflow fault is ; not initiated. ; ; POLYD Result: 0000000000008000 VAX-12 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors ; ~ ; Result defined ; by VAX Architecture: 0000000000000000 ; ~ ; .TITLE Test_Poly_1 .ENTRY Test_Poly_1,^M ; POLYD Arg, Degr, Table ; Instruction RET Arg: .QUAD ^X08FC91EF1F470101 ; Argument = +5.929E-39 Degr: .LONG ^X00000001 ; Degree of 1 Table: .QUAD ^XAA4EAAAAAAAABFAA ; C1 = -.33333333333333269E+00 .QUAD ^X0000000000000000 ; C0 = 0.00000000000000000E+00 .END Test_Poly_1 ; ; ; The POLYD instruction in the program below includes the ; identical underflow condition, but the sign of the result ; would be positive if not for the underflow. Under these ; conditions, even if PSL is set, a floating underflow ; fault is not initiated. R0 is left containing a zero, ; which would be correct had PSL not been set. ; .TITLE Test_Poly_2 .ENTRY Test_Poly_2,^M ; POLYD Arg, Degr, Table ; Instruction RET Arg: .QUAD ^X08FC91EF1F478101 ; Argument = -5.929E-39 Degr: .LONG ^X00000001 ; Degree of 1 Table: .QUAD ^XAA4EAAAAAAAABFAA ; C1 = -.33333333333333269E+00 .QUAD ^X0000000000000000 ; C0 = 0.00000000000000000E+00 .END Test_Poly_2 CASE 2 Example: VAX-13 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors ; The following POLYD instruction will produce an extra ; round when run on a VAX-11/780 or VAX-11/785 with an FPA. ; ; ; POLYD result: ^XAB31BBBBBBBBDFBB ; ~ (-0.67638061603596269E+19) ; ; Result defined ; by VAX Architecture: ^XAB30BBBBBBBBDFBB ; ~ (-0.67638061603596268E+19) ; ; .TITLE Test_Poly_3 .ENTRY Test_Poly_3, ^M ; POLYD Arg, Degr, Tabl RET ; Arg: .QUAD ^XFBFFFFFFFFFF407F ; Argument = ; +0.99999999999998578E+00 Degr: .LONG ^X00000003 ; Degree of 3 Tabl: .QUAD ^X666666666666E066 ; -0.16602069666338596E+20 .QUAD ^X3333333333336033 ; +0.12912720851596686E+20 .QUAD ^X444444444444E044 ; -0.14142503789843990E+20 .QUAD ^X9999999999996019 ; +0.11068046444225731E+20 .END Test_Poly_3 CASE 3 Example: ; The following POLYF instruction, when run on a ; VAX-11/780 or VAX-11/785 with an FPA will generate ; extra rounds. ; ; POLYF Result: FE8B766C (+.300425958E+33) ; ~ ; Result defined by : VAX Architecture: FE8A766C (+.300425977E+33) ; ~ .TITLE Test_Poly_4 .ENTRY Test_Poly_4,^M POLYF Arg, Degr, Tabl ; Instruction VAX-14 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors RET Arg:: .LONG ^X00004040 ; Argument = ; -0.7500000E+00 Degr:: .LONG ^X00000001 ; Degree of 1 Tabl:: .LONG ^XFF07769D ; C1 = +0.4005679E+33 .LONG ^X01018A01 ; C0 = -0.15528282E-32 .END Test_poly_4 CASE 4 Example: ; The following POLYF instruction, when run on a ; VAX-11/780 or VAX-11/785 with an FPA, will maintain ; too many bits during the truncation step. ; ; POLYF Result: 00811C7F (2.1093268E-22) ; ~ ; Result defined by ; VAX Architecture: 00801C7F (2.1093267E-22) ; ~ .TITLE Test_Poly_4,0 .ENTRY Test_Poly_4,^M POLYF Arg, Degr, Tabl RET Arg:: .LONG ^X0000407F ; Argument= ; +0.9960938E00 Degr:: .LONG ^X00000001 ; Degree of 1 Tabl:: .LONG ^X01029FFF ; C1=-2.6999592E-20 .LONG ^X00012000 ; C0=+2.7105058E-20 .END Test_Poly_4 Resolution: Currently there are no plans to supply fixes for these cases. Digital Software, including VMS and the Run Time Library (RTL) does not generate any of these problems. All of VAX-15 PAGESWAPPER - January 1988 - Volume 9 Number 6 POLYD and POLYF Errors on 78x Processors these items have been reviewed by representatives from Digital's software and hardware groups involved in the development of VAX products. The members of these groups have concluded that since these errors only occur with the above obscure data conditions and since correctly written programs should not normally depend on the least significant bits of floating point results, they will have a minimal effect on user programs. VAX-16 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session DECUS European VAX SIG Question and Answer Session DECUS European Symposium Rome 1987 Transcription provided by Tony Arnold, Chairman SPEAKERS: (All from Digital Central Engineering) Vernes Fisher Greg Eitzman C W Hobbs Stu Farnham Dave Solomon David Nunnerley John Halliburton B G Herbison TONY ARNOLD, University of Manchester, ENGLAND While you are all thinking of questions, we did collect some earlier in the week and to start us off I will read these out. If we have difficulty understanding them, maybe the person who asked the question can clarify it. Before that, the panel will introduce themselves and their area of expertise. I am Vernes Fisher. I am working on the DEC windows program right now. I can also answer questions about the VAX workstation software otherwise known as UIS. I am the odd man out here. I am Greg Eitzman. I work for HSC Engineering instead of VMS Engineering. I can answer questions about storage products and HSC. C W Hobbs. I work in VAX cluster development. I am also the designated file system representative. I am Stu Farnham. I can answer general executive questions and multi-processing questions. I am Dave Solomon, primarily representing RMS and RMS journaling. VAX-17 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I am David Nunnerly, working for the Product Management Group and I can answer the non-technical questions. John Halliburton, VMS Performance Group. B G Herbison. VMS Security Systems, not related to VMS at all. TONY ARNOLD Okay, I have some questions here while you are all scribbling your own. 1. A. VAN DER STAL, MINISTRY OF DEFENCE, HOLLAND When will true load balancing be put into LATCP. This means looking at CPU capacity workload, I/O activity etc. PANEL MEMBER We have looked into doing various forms of load balancing with LAT and we have produced incremental improvements along the way. It is difficult to decide really what the right algorithm is. For example, disc I/O, if you imagine being in a cluster, where this is useful! Who cares what the disc I/O is, because it is all one common set of discs? So you cannot use the I/O workload. We have done some work changing the CPU ratings and will probably continue to do so in the future, so if you have any ideas as to how to rate CPUs, let us know because we are running out of ideas how to do this! TONY ARNOLD I am just writing down that you are running out of ideas! (laughter). You must take any remarks I make totally tongue in cheek, by the way! 2. MR PALMER, HASSLE In which major release will DISKQUOTA keep correct information? (laughter) This implies it is currently incorrect. VAX-18 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER I would imagine that this question should be reserved for Keith. There are apparently bugs. If you have been submitting SPRs, that is the way to make these problems known to us. I am sorry but I am just not familiar with problems with incorrect disc quotas. TONY ARNOLD Okay, we will leave that one until later. 3. G H GAASBEEK, AGRICULTURAL UNIVERSITY, NETHERLANDS Now, someone has obviously got information that no one else has: he refers to a Version 5 here of VMS. VMS 5.X - I think they mean the next future release - the biggest topic seems to be multiprocessor usage. What is the percentage overhead to single processor users? PANEL MEMBER It is much too early for us to give you a percentage overhead imposed by multiprocessing on unit processor users. It is simply too early in the development cycle. I would like to say that we have done as much as possible to minimize the overhead imposed on unit processor users in two areas. The chief changes for multi-processing involve changes to the way that VMS is synchronized. There is a SYSGEN parameter, which allows you to select a piece of loadable code which performs all of the synchronization, which imposes effectively no overhead on unit processor users. For other areas of the exec where there are code paths that need be executed only on multiprocessors, there is a single BIT that describes whether or not the system is currently operating as a multiprocessor, and you will see probably thousands of branch-on-low-bit-clear instructions, testing the sense of that bit and branching around the multiprocessor on the code. So, there will be some very small amount of overhead, which I wouldn't attempt to quantify any further at this time. VAX-19 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session TONY ARNOLD Thank you very much. Right, another guy with advance information. 4. FRED E. CUSTOMER C.E.D., INC. Will the null driver on Version 5 on an SMP system support simultaneous I/O operations from each processor? PANEL MEMBER That actually is a hardware question. Bob Willard, will you tell us about the enhancements to null type devices. BOB WILLARD I am hard put to take this question seriously, but the real answer is, yes, of course. 5. ULRICH STRATHMEIER, BEB Will Digital continue to provide a driver to be used for RAM disc, currently PDDRIVER with VMS 4.5. PANEL MEMBER The PD driver is an integral part of stand alone backups operation from a tape drive. We don't foresee changing it. There are no guarantees. 6. MR COLLING, BP PET DEV. When will the limit of 16 of the number of nodes on a CI VAX cluster be increased? PANEL MEMBER Let's just say that we do recognize the need to go beyond 16 nodes and at some point in the future we would like to be able to offer this and leave it at that. VAX-20 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session Lots of mumbling - okay, we're going to stay with that answer! 7. ALFRED WULLSCHLEGER, FIDES ZURICH This is not actually phrased as a question. I need documentation on the implementation of a terminal device driver connected to the class driver. Is that documentation going to appear at some time? PANEL MEMBER That question has been asked of us several times this week, and the short and simple answer is NO. There are a certain set of interfaces which are designed for public consumption and to which VMS has committed. If we documented every interface within the operating system we would never be able to change VMS again. This is one of those interfaces which we reserve for our own use. 8. PER MAGNUS BANCK, GREATER STOCKHOLM LOCAL TRANSIT AUTHORITY How do I, from an image, get hold of the same information that I can get at DCL level, namely - F$ENVIRONMENT (PROCEDURE), i.e., "In which command file am I at present?" PANEL MEMBER In a special-purpose case, if you had control of the command procedures that are running, you could use the F$ENVIRONMENT lexical function to store the procedure name in a DCL symbol. There is an RTL routine which is LIB$GET_SYMBOL which can be used by your image to extract the value of the DCL symbol. In a general purpose case, where you don't have control of the command procedures, this is not currently provided. It is internal state inside of DCL, so there is no safe way of extracting the information. The answer is a no. 9. VAX-21 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I have three more which were actually submitted as SIR's on the machine. How many people have actually submitted an SIR? Three. Is it not possible to deliver VMS for the VAX-11/780 or others with an option to minimize the amount of storage on a disc for programs, driver, help, etc., you don't need or don't use, a kind of AUTOGEN for programs? PANEL MEMBER In a future release of VMS we will be modifying the tailoring facility to give you much of this functionality. What we will do is give you the option of removing pieces of VMS where pieces are defined in groups of images that perform certain functions. Those of you who are familiar with the MicroVMS would know these as, for example, the UTIL option, PROG DEVELOPMENT options. So there will be the ability for you to remove pieces of VMS in a supported fashion, and to restore those pieces if you later decide that you needed them. 10. GER VAN DIEPEN, RADIOSTERRENWACHT, NETHERLANDS Currently, there is no way to force help-paging, similar to HELP/PAGE from LBR$OUTPUT_HELP. A flag like LBR$_HELP would be convenient. Any plans for this kind of function in LBR? PANEL MEMBER I believe that LBR$OUTPUT_HELP has an optional parameter for a user-written output routine. We will take this under advisement. It does seem like a useful general feature but in the meantime, writing your own output routine that does pagination would be a way to go. 11. GER VAN DIEPEN, RADIOSTERRENWACHT, NETHERLANDS VAX-22 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session From a user program, you can only recall the last line typed. However DEBUG has the ability of recalling more lines, so I assume there is a more or less standard way in VMS for doing this. Please make this available as a callable interface. PANEL MEMBER This is currently available. The name of the routine is SMG$READ_COMPOSED_LINE and it is documented in the SMG Facility, the screen management facility. Virtually all of the VMS utilities which do multi-line recall are using this function. TONY ARNOLD How much more of SMG do you have to set up to use that? Do you have to establish a virtual keyboard.... PANEL MEMBER Greg - the hardware expert - says that you need a paste-board and a keyboard. GREG You need about two extra variables and two extra procedure calls and then replace the calls to whatever - GET LINE - with the SMG call, so it is really very simple. 12. ALAN REID UK ATOMIC ENERGY AUTHORITY When users log on they can be forced through SYLOGIN.COM. Are there any plans for doing an equivalent at logout stage i.e. SYLOGOUT.COM> If not can it be done in the future? PANEL MEMBER The traditional problem with forcing execution of a LOGOUT.COM, whether it be a system-wide logout procedure or a user-specified logout procedure is that, for the sorts of applications which I have had described to me, that logout processing must occur under all types of process termination, either voluntary via logout, or involuntary for example, via a VAX-23 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session non-fatal bug check and the complication in the solution is guaranteeing that the logout mechanism cannot be circumvented. ALAN REID Why not simplify it and just give us part of the answer and just take a normal logout rather than a bug check. PANEL MEMBER Because when we come to DECUS with incomplete solutions, we meet with angry customers. We would rather do a complete solution now than do an incomplete solution which would constrain our ability to come up with a complete solution later. We do perceive a genuine need to provide user-specified image terminations or process termination processing, so it is a hard problem but one that we have heard about before and we are looking for solutions. PANEL MEMBER Can we just do a quick poll here? How many people would personally find a real use if logout simply had the ability to invoke a command procedure that could be bypassed by STOP/ID, bypassed by DELETE PROCESS? PANEL MEMBER We are now going to circulate a petition and all those who raise their hands will have to sign it, agreeing not to complain when we implement such a process! 13. In a large cluster with several machines and large memories, discs seem to run out creating dump files. Under what conditions is it safe, or reasonably safe, to use common dump files in a large cluster? PANEL MEMBER VAX-24 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session When you are not concerned about saving your dumps! (laughter and applause) I am sorry, that was flippant! I think that Stu can address the real issue. STU FARNHAM Let me first say that any problems that any customer has ever had with dump files we in VMS Engineering have probably a thousand fold, because the software that we are running is under development and many times less stable than that issued to customers. So we are acutely aware of the problem and there are several things being done both in the short term and the long term to address this. We do not believe that the solution is to provide a common dump file. We believe that the solution is actually comprised of several components. The first is to reduce the size of the dump file, and there are algorithms currently being tested to allow subsetting the portions of memory to get rid of the dump file. Beyond that I believe, this is one engineer speaking, not Digital as a whole, that we have to be able to move the dump files off the system disc. We have got to address that in a longer term than the next major release. So there is quite a bit of movement there. PANEL MEMBER But, to answer your original question, if you put your dump file in SYS$COMMON and [SYSEXE] and make it large enough for the largest machine in your cluster, it is perfectly safe as long as only one machine crashes. You can copy the dump out of the dump file. The only risk that you take is, if the second machine crashes before you get the first dump copied, you will potentially lose that. I have seen people who run with one bit dump file: when a machine starts acting up, they will temporarily create a dump file for that one machine in SYS$SPECIFIC to protect themselves. PANEL MEMBER VAX-25 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session You can use the page file, but it must be the primary page file, which must be on the system disc, so you are sort of back in the original position, which is that you have to have a file on the system disc which is large enough to accommodate the crash dump of every CPU, so it is not much of a solution for the VAX cluster problem. PANEL MEMBER Bear in mind the fact that this primary page file puts paging I/O on your system disc and as of V4.6 you no longer have to have a paging or swap file on the system disc at all, so that you can put all of the paging and swapping onto other spindles. 14. BRIAN MATTEY AIRCALL TELETEXT UK Does the new Micro-VAX chip still implement CRC in software and what instructions have been moved to microcode and vice versa? Specifically, is CRC now in there? PANEL MEMBER Is there a micro-VAX person in the house? I am sorry, we can't answer that now. Could you ask at 9.00. 15. One known bug of DECnet is that after a power failure of the machine, if the machine attempts and succeeds in normal start, DECnet doesn't start again. When do you think this problem will be solved? PANEL MEMBER Have you SPR'd that and gotten an answer that was unsatisfactory? QUESTIONER VAX-26 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I have read that this has been stated in a recent dispatch so I would hope that it would not be necessary to submit an SPR for this problem. PANEL MEMBER What was the response in the dispatch? That sounds like a problem we will be addressing. QUESTIONER The response was in a recent release, but I was hoping for something a little more precise. PANEL MEMBER Okay, I am sorry, we do not have any DECnet people here. Ask again at 9.00 when we have the phone link operating. When you ask again specify the type of hardware you are using. 16. W. MICHAEL TERENYI, SANXOZ RESEARCH INSTITUTE, VIENNA Why does SHUTDOWN.COM stop first the queue manager and then invoke the site specific shutdown procedure? All batch oriented software (i.e. message router) has therefore to be stopped separately and not within the site specific procedure. PANEL MEMBER The question is why does SHUTDOWN.COM stop the queue manager before calling the site-specific SHUTDOWN.COM. That is an interesting question. I would imagine that the reason it's that way was that when the stop queue manager command was put into shutdown, whoever put it in there decided that the reasonable place during shutdown to stop the queue manager was here, and down there was where we did the other stuff. That is an interesting thing for us to take back, because part of our work now is looking at how we are doing startups. We are certainly amenable to changing a shutdown procedure, although it sounds like a general case might mean that we have to have several different callouts for different stages of shutdown. Thanks for bringing it to our attention. VAX-27 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session MICHAEL TERENYI There is a lot of software now based on batch jobs and it should be stopped in a correct manner. 17. CLAES ENGELIN, OPIAB KONSULT, SWEDEN I have a question regarding DECnet and security, if we can mention those two things in the same breath. Will we ever here in Europe see end-to-end or point-to-point encryption over DECnet? The background is that a lot of companies, especially defense contractors, are getting a lot of heat from defense authorities about security in their systems and about information going across wires all over the place, in clear, so to say, including things like usernames and passwords. Anybody with a little PC can hook up to that net and just eavesdrop. It is really a potential disaster area. PANEL MEMBER As I mentioned in my talks earlier this week, some work has been done in looking at end-to-end encryption DECnet. It is a long way away, possibly even longer in Europe because of U.S. export controls, which we can try to work around but can't really change. For DECnet there are two sorts of problems. There is all the networks in your systems and your wide area links. The wide area links can be protected using link encryption devices. There are some available, and they wouldn't affect the PC on your own site tapping in, but they would affect the information that is going across the country. That may be a partial solution for you. I have heard it several times this week, and we are taking back the message that we need to do more work on DECnet end-to-end encryption, but this can take a long time to develop. CLAES ENGELIN You say that the problem is two-fold but one thing is encrypting in the first place, the other part is due to restrictions on exporting algorithms. Would it make it easier if you could use a simpler encrypting algorithm that would possibly be easier to crack, something like some algorithm that would only take a year or two to VAX-28 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session crack instead of ...? PANEL MEMBER Captain Midnight secret decoder rings are export controlled. It would still take a license to send it out. The weakness might make it easier to get a license but doesn't eliminate the need for a license. There is also the option of building something that uses an encryption chip, selling it to you without the encryption chip and having you input your own. An encryption device without the encryption chip is still export controlled. There are no easy ways around this. CLAES ENGELIN Not until after the next election, right? PANEL MEMBER Maybe! 18. MR MONTANA, PRACTICAL PHYSICS INSTITUTE, SPAIN Is there any intention of providing in future releases of VMS some kind of checkpoint restart facility? PANEL MEMBER We have basically re-evaluated the checkpoint restart effort that we did have underway a couple of years ago. We took a look at it and decided that the product we were going to deliver really fell short of what most people wanted. The original ideas behind checkpoint restart was to allow you to protect your computation investment over a long period of time. We tend to think that it may not necessarily be the product that people really want. What people really want is the ability to keep applications running in the cluster environment and be able to recover dynamically from those kind of failures. So I would think it very likely that we will put our efforts elsewhere and not actually deliver a checkpoint restart for a while. VAX-29 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session 19. MR LOWER, VOLKSWAGEN We find it necessary to get the server name on the server port within the systems office. Is this possible in the next version? At the moment we must start LATCP, there is no possibility of change of privilege to do this. We need this for security reasons. PANEL MEMBER Someone in the audience confirmed it. All we can remember is that it is returned by $GETDVI in VMS 4.6 or the next major release. We have confirmation that it is in VMS 4.6. PANEL MEMBER The question was getting the LAT server name from a line. In VMS 4.6 I believe it is available as a queue I/O modifier. In a future release it will be a $GETDVI item. 20. PAUL MANEKMAN, AUSTRALIA We have connected our VAX to a foreign device and it seems to expand the vertical tabs, carriage return and several line feeds. We have set -/NOTAB and /TAB and it seems to make no difference. We want "XON/XOFF" processing, so we couldn't set /PASSALL. Is there a way around it? PANEL MEMBER The suggested solution of setting PASSALL is not acceptable because XON/XOFF processing is required. PANEL MEMBER What came to my mind quickly was /FORM /NOFORM not /TAB /NOTAB. That I believe affects only form feed, I am not sure if it affects vertical tab as well. /TAB /NOTAB affects the horizontal tab, not the vertical. /FORM /NOFORM affects form feed. So that is all I can recall on that. I suggest holding the question until 9.00. VAX-30 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session SUGGESTION FROM AUDIENCE PANEL MEMBER RESPONSE /PASSTHRU is /PASSALL but it honours XON/XOFF so it is sort of /PASSMOST! (laughter) 21. BEN KNOX, PFIZER CENTRAL RESEARCH, ENGLAND In a large cluster we have an increasing need for cluster-wide commands to simplify operation and management, i.e., SHOW USER; SHOW DEVICE/FILES. Do you have plans to provide such facilities? PANEL MEMBER Making a cluster seem more like a single system is a direction in which we are heading. In a future release we will have a system management facility which will allow you to execute specified DCL commands on remote nodes in your cluster if you have the appropriate privileges. Over the course of the next major releases you will see continued improvements in the 'system-ness' of the cluster. We are definitely aware of the rough edges, if you will, in the seamlessness of the environment and we are working on them. TONY ARNOLD You have to SET TIME/CLUSTER on VMS 4.6. B. S. KNOX Added to answer summary: another user later suggested using a DECUS tape facility called FINGER to achieve this sort of result. 22. ANDERS WAHLBERG, PHARMACIA AB SWEDEN As a Systems Manager, life is sometimes quite hard and one thing that I have recognized is that after a period of time the SYS$SYSTEM area on your system is normally very cluttered up. I wonder why Digital have not used the possibility of creating sub-directories to put the various systems and products into, thus enabling you to ease the structured material. VAX-31 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER In a future release we have made a couple of leaps in that direction. We will have a separate directory for system loadable images, such as drivers and the remaining pieces of SYS. I alluded earlier to improvements in system startup - we are creating a special directory to contain the pieces of the system such as layered product command files, which are used during the startup of the system. We have no particular plans right now to make immediate additional split-ups but we are going in that direction. 23. JEAN PIERRE-PETIT, ESME, FRANCE Is it possible to provide a system service that instructs the Job Controller to put each accounting record in a mailbox as well as in ACCOUNTNG.DAT (similar to what exists for ERRFMT)? PANEL MEMBER First of all I would like to ask a question in reverse. What part of the application can't deal with processing the accounting file? JEAN PIERRE-PETIT The application can't process the accounting record on a real time basis and increment the budget each time a process is terminated. PANEL MEMBER I would say we will have to take this one as a suggestion. You can redirect the accounting logical name to send the records to, if you want, a mailbox, but then they are not going to go in the accounting file, unless you put them there yourself. I understand what you are asking for and we will take that down as a suggestion. 24. MR CAMBRIEN, PHILIPS, SWEDEN VAX-32 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session During system shutdown, why must I wait for the system to write out the whole memory to the dump file? TONY ARNOLD I guess the question is why do you dump all of memory on a normal system shutdown. PANEL MEMBER There is some confusion up here as to whether we are doing it immediately or not, but if it is not corrected in the next major release, it will be corrected shortly thereafter. PANEL MEMBER It will not NOT be done. (laughter) PANEL MEMBER It has always annoyed me as well, and we will take care of it. MR CAMBRIEN It takes several minutes with 56 megabytes. PANEL MEMBER Yes. And when you have 256 megabytes it takes considerably longer! 25. PAUL TEECE, GOLDMAN SACHS, U.K. Will TSM work on an 8700 under VMS 4.6? We are having problems with the debnt. We have lots of protocols running, like DECnet, XNS and so on. The terminal server manager can't talk back to the terminal servers. If we switch off DECnet it works fine, but we need DECnet. The release notes actually say that it is fixed in 4.6, but our local office tells us that it is a hardware problem on the DEBNT. VAX-33 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER BOB WILLARD, VAX HARDWARE ENGINEERING I don't specifically know what the problem is that you are referring to but we have started a program of rolling upgrades from the DEBNT to the DEBNA all one product that will take place over several months. PAUL TEECE We have had our DEBNAs all taken up to the latest revision level, because we had other problems with them. PANEL MEMBER This is a replacement product. Thank you. That is a piece of information I needed. This is a replacement, a new design, and there would be a completely new board. PAUL TEECE It was a new board. They took everything out and took us up to the latest revision level. PANEL MEMBER Was it a DEBNA. BOB WILLARD Was it a T10 24. PAUL TEECE It is yes. As used at Bankers Trust. They told us to insist on the upgrade. BOB WILLARD This would be a question for the DECnet people at 9 o'clock. It is obviously not a hardware problem, right? If the 10 24 is good. VAX-34 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PAUL TEECE The message is actually "INSUFFICIENT VIRTUAL MEMORY" and it is because the mock protocol can't get in there. PANEL MEMBER Please, we have no DECnet expertise in the panel. Ask at 9 o'clock. 26. ERICSON TELECOM, SWEDEN You can have the qualifier /NODEBUG back in RUN command but if you want to run without DEBUG when you SET COMMAND, you can't do it. There is no such qualifier, with SET COMMAND, no DEBUG. There is no possible way to make /NODEBUG, when you SET COMMAND. PANEL MEMBER What you want to do is to change the default for the running command so that .. QUESTIONER No running command but SET COMMAND with new qualifier, /NODEBUG, for instance. PANEL MEMBER I would like to introduce Brad Becker, the owner of VAX DEBUG. (applause). BRAD BECKER I would like to make sure I have the question correctly, since I couldn't hear very well. QUESTIONER I would like to run the image which was linked with /DEBUG, but run without /DEBUG when I use the SET COMMAND. VAX-35 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session BRAD BECKER There is a bit in the image header ... next question! Well, that is one alternative, the other is to have two different commands. QUESTIONER Yes. The question was also that there is no possibility of putting KEYWORD /NODEBUG, there is no such keyword in the CLD file either. PANEL MEMBER We are looking for a CLD expert. PANEL MEMBER I have an answer or trick for the first question. Write a simple command file containing two commands, the GO and the EXIT command, and define the logical name DEBUG$INIT or DEBUG$INPUT to point to a command procedure which just does a GO. There is also DEBUG$INIT which I guess would do the same thing. 27. SAME QUESTIONER You can imagine the command procedure containing only two commands. The first one is SET COMMAND and then FILENAME and the second name is, start with "!" and some text. And what happens, you have the warning message - IMAGE DATA DCL-W S COPY DIT or something. When I put the "$" sign before "!" sign it works, but maybe there is some mistake in my CLD file. PANEL MEMBER The reason for the DATA SKIPPED message is that when DCL finds text in a command procedure that has not been read in by a program through SYS$INPUT and you are now at the end of the command file, in essence there are commands or lines in there that nobody has processed, so it is telling you that it has skipped something that no image had read in those commands. If SET COMMAND had read in that line then the error message would not VAX-36 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session have appeared. BOB WILLARD This only happens after you execute an image in the command procedure. So you might see that if your command had invoked an image. With a "$" sign you would not see the message. So if it seems inconsistent, it is only when the first line after the image exits, that the DCL finds a malformed command. 28. ANDERS WAHLBERG, PHARMACIA, SWEDEN On to a more policy-oriented question. (laughter) Now and then there develop important and very dangerous security leaks in VMS. Somehow Digital mistrusts us systems managers and they do not tell us anything before they have distributed a fix for the problem. On the other hand, we have found through hard practical experience that some of that stuff goes out to other people through other channels, and I can imagine that there can be problems in distributing information about exactly what is amiss in the system, but I would like very much to have some sort of a warning flash sent out - "Please watch your systems a little more". There are a lot of interesting security features that you can enable but which you normally don't have enabled due to the sheer mass of information that is on your system. It would be nice to have some sort of procedure to alert the systems manager that something is amiss, then we could be really looking into what is happening. I have, from personal experience; found that sometimes the distribution of the fix gets delayed or misplaced or something like that, and I hear (several months after the rest of the community) about the fix to the operating system. Then I have a feeling that we have lived fairly dangerous lives in the meantime. I understand that this is a problem, but this is a problem for both of us, and you have to do something about it. AUDIENCE MEMBER ALAN SILVERMAN, CERN, GENEVA VAX-37 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I was going to give you a chance to reply first but I will follow up what Anders has just said, because it was going to be in my question. Earlier this year we became aware of the fact that there was a security hole in VMS, which was open with 4.4, open for at least six months. I believe it was talked about in U.S. DECUS last October but I am not sure. We became aware of it because we got a message in UNIX NEWS of all places, and we also became aware that there was a patch available, which we also read in UNIX NEWS. We phoned our local support office, and we got that patch after some explanation: they didn't believe us, but they found it, and they got the patch. We then applied that patch. Other colleagues on my site got that patch two months later. That bug opened VMS wide for six months and you didn't do much about it to help us. You had a patch which you did not distribute and we, the users, should insist that security patches should be distributed immediately as a high priority update, without having to wait for the next version of VMS. (applause). PANEL MEMBER Well, we agree with you. (laughter) No this is not a funny subject, this is very serious. We apparently did not distribute this patch as quickly as you would have liked. However, we also have to recognize that it was very important that we got it right so that we could distribute it correctly. Our policy in general is not to discuss the specific details of security breaches; obviously that would be a silly thing to do. We do put a very great emphasis on a security problem when it is reported, we do make them a very high priority. We attempt to get a solution as quickly as possible and distribute it as quickly as possible. Apparently this hasn't happened as quickly as we could have done it. In the future we will do this better and we would like to work with you. If you could help us with ways in which we could get this kind of patch out in a faster, more efficient way. Let me just say that in some ways it can be very difficult for us, because if you are under a warranty contract, software support, it is relatively easy for us to get hold of you. If you are not under warranty, then we don't have you in our database. It can be a very difficult problem. We don't want to announce to the world what the problem VAX-38 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session is. So we would like to work with you to try to improve the process, but yes, it is very serious. It is a very high priority to get the fix out, when the fix is appropriate, which is not always the case. TONY ARNOLD It has actually gone 9 o'clock, which is when we are going to attempt to make contact with the people in the States, so if you will just bear with me for a while. LOCAL QUESTIONS TAKEN DUE TO DIFFICULTIES WITH TELEPHONE LINK UP. PANEL MEMBER Somebody asked earlier about callable COPY. It is fairly easy with RMS and some logical QIOs to get an exact copy of an RMS file fairly reliably. I don't know if anyone has any comments on that. PANEL MEMBER Well there is some reason for the reams of code inside the COPY utility I think. There is more to it than just copying the blocks. There is propagating various file attributes. PANEL MEMBER Yes, you have to use RMS to get things like the end blocks, you have to create the new file with all the basic RMS information, then you can go to USER FILE OPEN and use QIOs to move the actual data. PANEL MEMBER That will work on a file that is local and on disc but not for other unit record devices. The COPY utility is designed to handle all kinds of input devices, terminals, mailboxes, remote files, as well as local files. So yes, for disc-to-disc copy, creating the file and QIOing the blocks in will work fine but COPY is a general utility. VAX-39 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER I just did it on one disc, disc to disc on a MicroVAX. PANEL MEMBER As far as callable utilities, we mentioned it is something we want to provide more of. In fact, there is a strong, unwritten rule that at this point there is no reason that new things should not be provided in both callable form and utility form. It is going to take time to go back to the other utilities. 29. PER MAGNUS BANCK, GREATER STOCKHOLM TRANSPORT It was very nice of Digital to allow us to build our own print symbionts. About half a year later there arrives the LAT print symbiont, which we cannot modify, for example, for site specific job pages, which is the case with us. Is it possible to get the object modules or anything like it so that we can link our own LAT symbiont? PANEL MEMBER Is there anyone in network communications here who can answer that? No. Okay, we will have to note that. 30. PAUL TEECE, GOLDMAN SACHS UK Can you give me any indication when the problems between the 8700 and the PRO 380 will be finally fixed? PANEL MEMBER To the best of my knowledge the problems that we have heard about in VMS between the 8700 and the PRO consoles have been fixed. Whether the fixes have been distributed to the field or not is a problem that is out of the hands of anyone in this room. PAUL TEECE VAX-40 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session We thought REV E of the console software would fix it, and it didn't. Are we waiting for a new device driver for the CW driver in VMS 4.6 or is it further away than that? PANEL MEMBER Which particular problem are you talking about? PAUL TEECE There are a lot, but the one that causes the most problems is when you try to use the console to edit a boot file or look at the log file that lives on the console, if a message comes through from OPCOM or something like that, it will crash VMS. PANEL MEMBER I think the answer that it is fixed is right. It sounds like a European software distribution question though. PAUL TEECE I got an answer from an SPR that said they didn't know when it would be fixed. I just wondered if you had any more information. TONY ARNOLD It sounds like it is fixed but it has not reached Europe yet, is what you are saying. PANEL MEMBER It does not happen to us with our 8700s in our labs, with the revisions that we have and with VMS version 4.6 and later. I don't know which piece is giving the problem. I believe that it is probably a difficulty in getting the fixes from the engineering sites, through all the levels of reproduction and distribution, across the Atlantic to Europe. VAX-41 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session 31. W. MICHAEL TERENYI, SANDOZ RESEARCH INSTITUTE Are there any changes in VMS memory management coming? I have heard that SWAP OUT PAGE COUNT can be removed or disabled. PANEL MEMBER I don't believe that SWAP OUT PAGE COUNT has been eliminated. There are some changes in memory management. A couple that come to mind for the next major release are extensions to the limits on working set size and modifications to the algorithms for allocation of SWAP FILE space. Those are certainly running in prototype forms in our labs and we hope that they will be in the next major release. QUESTIONER So there is no change in the basic algorithm? PANEL MEMBER At the very primitive level, no, but for example, as SWAP FILE space is currently allocated - you need a SWAP FILE that is large enough to handle every resident process, every created process rather - and the SWAP FILE space allocation is done at process creation which looks pretty silly from this end, now that we have gone to algorithms which allow dynamic allocation of SWAP FILE space. 32. W. MICHAEL TERENYI Can I have an RMS question? Is there any intention of changing "contiguous best try" to "contiguous best fit"? PANEL MEMBER That is not on the immediate future or further future list - NOTED. VAX-42 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session 33. A. BARTUR, ISRAEL, DECUS Do you think that the current VMS priority scheme can cope with a large number of interactive users with different service goals? PANEL MEMBER No. There are a number of changes, especially in high-end systems for which we have to make extensive modifications to VMS scheduling code. However, the problems introduced by large end systems need to be dealt with comprehensively and not point-by-point. As such we don't expect to be doing anything in terms of major changes to scheduling algorithms in VMS version 5. However shortly thereafter... . 34. BOB SIBLEY, ELIDA GIBBS LIMITED Will we ever have the ability to run multiple versions of VMS on the same CPU - in order to test VMS releases? TONY ARNOLD Simultaneously? BOB SIBLEY Yes. I was thinking of something like the VM for the IBM system, a monitor which runs under, to allow you to run multiple versions and test one version of VMS while the other one is doing the live work. PANEL MEMBER The answer is no. 35. Suppose I buy RALLY and write an application using it, and I change all the cryptic elaborations to my national language. Sooner or later the worker using it will make a mistake. Up comes an error message in English, and off goes the worker and you don't get your .... VAX-43 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session Do you have any comment? PANEL MEMBER My comment is that I believe that the earlier answers to this sort of question pertain. There is currently a shortcoming in VMS's handling of multiple languages and international character sets. The shortcomings are being addressed, one might say in a more piecemeal fashion than not, but we are making progress in that area and we intend to continue making progress in that area. AUDIENCE MEMBER Does that mean you are going to translate the error messages or are you going to open up the interface so that we can translate them for you? PANEL MEMBER I believe that at VMS, some of that is done already. There are local language versions of at least parts of VMS. TONY ARNOLD Does anyone know of local language versions of VMS? (Faint mumbling, unable to hear.) AUDIENCE MEMBER There is a product called LOLA - this product is available in French and German. It provides a system error message in those languages and some help files. The main problem is the time between the delivery of the U.S. version and the European version of VMS, a matter of 2 or 3 months for translation purposes and availability of the product. PANEL MEMBER VAX-44 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I assume you have commented on the MCS capabilities in the future. In other words, we are definitely moving in the right direction. I think it was Rally you were talking about. A layered product has yet another further layer of complexity since there may be.... It is a point-by-point issue with a layered product. We have started in the operating system to provide the tools, so that layered products can then use that facility to do the right translations more easily. We do recognise the need to get better in terms of handling foreign languages. It is an internationalisation program inside the Corporation, and I think in time you will see that we are going to get much better at it. PANEL MEMBER For example, in Version Future we have a new terminal facility for local character set translations; RMS, in the file system; and SORT supports international collating sequences and user-defined collating sequences. There is a new runtime library facility for local date translations. We are working on it. We realise it is a problem, but we can't fix every piece overnight. QUESTIONER I hope you realise that my users use different languages. When a user logs in I must ask which language he uses and then adapt the reply on that information. That is my application: I must always be able to respond in several different languages. PANEL MEMBER Part of our design centre for our internationalisation effort is to allow multiple languages to coexist on the machine. 36. BALTICA BANK, COPENHAGEN, DENMARK We are using RDB and PASCAL. We have been using the precompiler up to Version 2.3 of RDB and now we have got a new RDML precompiler. We just have one problem, we can't get past the first statement in our PASCAL code and it is an external database statement. VAX-45 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER You need to find Jay Feenan. He is the Central Engineering Database guy here. I don't think any of us here can answer that, as we are all VMS up here. PANEL MEMBER Congratulations. We did get through on the phone and they said that the problem of the power fail restart on DECnet is fixed in VMS 4.6. At which point there were so many people leaving we hung the phone up! 37. A question related to password expiration: VMS, when the password expires, gives the user a warning message and after presents the statement of the login command file in UAF. My problem is, will there be, in the future, a simple way to understand when the user password is near the expiration date and to permit the user to change the password with the normal VMS command SET PASSWORD? I mean something like a lexical function, F$GETUAI for example. At the moment it is necessary to run the image which performs the system service, SYS$GETUAI and after that pass the information to the common file that had been activated such as the login command file of the user. PANEL MEMBER There are changes in the $GETUAI, $SETUAI. I am afraid that we are not sure what exactly has been happening. One of the things that has been discussed is the idea of forcing a password change so that you can't log in without updating your password. About all I can say is that we are looking at things in that area because we realise there are problems. QUESTIONER VAX-46 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session There is SYS$GETUAI information so we would be grateful. 38. BRIAN MATTEY, AIRCORP TELETEXT, U.K. Will BACKUP support the new DFS things, such that you could backup remote files to a local tape? PANEL MEMBER We don't have enough familiarity with DFS to answer that. PANEL MEMBER There aren't any plans at the moment to merge BACKUP, or have BACKUP use a DFS protocol, that I know of anyway. That is not to say that it might not be a good thing to do one day. 39. ROGER BROOKS, UNIVERSITY OF LIVERPOOL, U.K. Can we have access to the password encryption algorithm? We have to authorise large groups of users, and we need to front-end authorise, if you like, by something that sets up disc quotas and generates directories, and in some case generates passwords, for a large group of users. It would be easy to do with $SETUAI, except that we need to incorporate the password. PANEL MEMBER I think the answer to the last question about $SETUAI applies. We don't know well enough what is happening in that area to answer the question. PANEL MEMBER You basically have the wrong set of expertise here to answer your question. We know that there has been a lot of movement in that area. We don't know if your particular request is among the many things that are being done. VAX-47 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER One area that we are addressing is the system management area in general. One of the big problems with that area is that a user is currently not a well defined object in VMS. There are several different utilities that you have to run to manipulate the object "user". We are making an effort to give you a coherent interface to manipulate system management objects. We will take your input. Do you want us to generate a password, or do you need to specify the password for your new users? R. BROOKS We would like to be able to specify the password. We can see the algorithm on the fiche, it is just that I think it is in PL1, and we don't have PL1 so it is a case of translating about 8 pages of PL1 for something that we have. PANEL MEMBER I think there is a situated problem in the suggestion, because the menu of a public password algorithm is to write a small program that generates a lot of passwords and groups them and compares to the encrypted version of UAF. That could be a big problem for you, especially if you don't use DELETE/ERASE each time you create a new version of your UAF and write some portion of it to the disc, as an encoded version of the password and try to generate passwords that give the same encrypted value. R BROOKS I thought that was supposed to be impossible because there are too many encrypted values, at least that is what DEC say, isn't it? PANEL MEMBER They will then give 64 bits for annexing, and there are different passwords, different strings of characters that will lead to the same result. Because the encrypting algorithm is not a unicorn, many strings could lead to the same encrypted value, so suffice you VAX-48 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session can generate one. PANEL MEMBER If you wish to try and break the password scheme by constantly generating new passwords, encrypting them and then matching them against the 64 bit value, then you have a 50% chance of succeeding after 5.7 trillion passwords are attempted. PANEL MEMBER I know some people at Digital who used this technique a few years ago and I know that since that date we have put in some redundant instruction that slows the algorithm in order to make it difficult to find the password in a reasonable time. But at the same time you increase the problem of your CPU's load and we are back to the same problem. R. BROOKS We would be quite happy to use this algorithm if you needed something like, with which you can do anything anyway. No way do we want this thing to be easily accessible, we just want it there. 40. OLIVER DAVIS, ALLEN COMPUTERS, U.K. I would quite like to be able to use $GETDVI with a wildcard, perhaps generic wildcard, so I could look at all DU devices and return characteristics for allof those. Is there any chance of having that in future releases? PANEL MEMBER It is something that we have looked at and will be doing some time in the future. It is not planned for the next release. OLIVER DAVIS VAX-49 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session Okay. The other one is $GETUAI. It is a real godsend now to see that. I would quite like to see that wildcarded too. Also you can't look up by UIC; you have to look up by specific username. Is there anything to be done about that? PANEL MEMBER We can't answer that. TONY ARNOLD I thought you could wildcard. I am sure I have written a program to do it. OLIVER DAVIS I don't think so in $GETUAI. TONY ARNOLD I think it is obscure, but I think you can do it. We can check the documentation in the clubroom tomorrow if you like. 41. Is there a special reason why the CONVERT utility does not accept wildcards, so you can convert several files to the same FDL file to another format? We have to change 200 files to a special format. I have to invoke CONVERT 200 times. PANEL MEMBER The main reason would be that it only has one output file. So you want wildcarded inputs and outputs. QUESTIONER Yes, to take a new version of every file. PANEL MEMBER VAX-50 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session Thank you, we will note that. 42. RICHARD BROOKS, R.P.R.L. I believe that when you invoke SHUTDOWN, in addition to shutting down the queue manager, it also shuts down DECnet immediately. It happens quite often that I want to tell the users a long time beforehand that the machine is going to shutdown, and I would like to use the network during that period of time, rather than having half an hour when the network couldn't be used before the shutdown. PANEL MEMBER It was my understanding that SHUTDOWN does a SET EXEC STATE SHUT approximately 5 minutes before the shutdown, and that the network is fully available until the last handful of minutes before shutdown. 43. I have a question about the tapes. I need the tapes by density, not by the name ( MT, MF, MPU and so on). When I mount the tape, sometimes I need 800 bpi, because we have some older tape drives. PANEL MEMBER I heard two different questions. Do you need to know the hardware device name for the device, or do you want to be able to determine the density of the tape? QUESTIONER No, I just want to be able to mount the tape with the specified 800 bpi. PANEL MEMBER Craig suggested you want to be able to say ALLOCATE and get one by density. VAX-51 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session That has been noted. QUESTIONER Is it possible to make all tapes on the cluster sort? PANEL MEMBER No. Currently there is no MSCP server for tapes and none in the near future. QUESTION FORTY TWO BEAT GODGLUCK, AUTOPHON AG, SWITZERLAND Why are the system parameter descriptions incomplete in the system manual? For example, I had to install a software package for a simulation. They demanded to augment CTL pages from 250 to 500. I never found them in any documentation. PANEL MEMBER Did you notice that it is in the SYSGEN online help, under "HELP parameter"? This has been noted for a version in the future and you should see complete descriptions of all of the SYSGEN parameters. We will note that, but I believe that we are going to make an attempt to document even the special parameters. BEAT Okay. The second question is for better system management, I would like to have an improved monitor (for example MONITOR PROCESS/FILES or MONITOR PROCESS/CHANNELS so I can locate where the I/OS goes, what files are open and so on). Is that possible? PANEL MEMBER VAX-52 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session Certainly you can use SDA to show a process with open files and say SHOW PROCESS/CHANNELS, and it will list the channels that are open and the files that are in use. That is not a MONITOR display but that is certainly one way to find out the information. BEAT I have a slight complication with SHOW DEVICE/FILE to search on every disc. If you have a lot of discs and if you open files over DECnet and so on, the job is overwhelming. PANEL MEMBER Right, we have had many requests for finding who has what files open on what disc and I suppose this is something that Dave is going to note, another one of our SIR requests. AUDIENCE MEMBER I must ask Digital people not to listen to me. You have an unsupported DCL command, SET WATCH FILE/CLASS= DECnet. With this command you can cause the XQP to log any file activity for your process. You must type the DCL command SET WATCH FILE with FILE in four letters. Do not use abbreviations, it doesn't work. Then you specify SET CLASS= MAJOR, you have major statistics to the file, but it is unsupported, and it could lead to very important CPU consumption. PANEL MEMBER Can I just make a last comment on that. I believe that what you have said is in at least one, if not two, of the SIR's that were voted on for this 87 European Top Five. There was an SIR that was very similar, if not the same, but it didn't make it into the top five, so we recognise that is something that is needed but have it as a suggestion. The rest of your colleagues did not give it the votes that were needed. VAX-53 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session QUESTION FORTY THREE AL WERBERG, INFN, TURIN At the present time CONVERT will not allow changing the carriage control characteristics of a file. Will this capability be implemented in the future? At least in those cases in which the dump of the file in the two formats is identical? PANEL MEMBER What you are saying is that the carriage control data in the file is not converted, is that the question? AL The problem arose when I had a file created by a FORTRAN program which could not be read by another program. I made a dump of the file that was created in a different way, and the two dumps were identical. The only difference was in the file description: one had FORTRAN carriage and the other had what I think was called UNDEFINED carriage control or something like that. The two dumps were identical. CONVERT would not allow me to switch the characteristic from one to the other, so I simply had to write another program which read the file and rewrote it. PANEL MEMBER I believe the problem here is that CONVERT will change the carriage control attributes as marked in the file header, but CONVERT does not modify the data in the file, so even though you have removed the FORTRAN carriage control, the first byte is still a space or a one or a minus sign or the various other FORTRAN carriage control. AL That is what I thought too. When I dumped the file and checked byte by byte, everything was identical. It states in the manual that it will not change, in any case, the carriage control characteristics. I didn't try EXCHANGE. VAX-54 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER What you want is for CONVERT to change the data in the file in this case. AL I am asking if this is foreseen. PANEL MEMBER This is not foreseen, no. 44. ANDERS WAHLBERG, PHARMACIA AB, SWEDEN Nowadays there is a very nice thing in VMS. VMS is told to install all the various layered products. How about providing a VMS DE-INSTALL to remove the layered products, (remove all the files that have been created in various directories)? Sometimes it can be quite a hard job to remove a product. (applause) PANEL MEMBER We are addressing that problem. It is part of the overall system management effort that we are making. Noted, thank you. ANDERS Another question we have with some systems is the changing account file for specific quotas or other things you need running. Are you going to add something to removing that sort of control information. Like byte limits and other parameters. We would like to have something like a list you can specify datatrieve...so you need... PANEL MEMBER For someone to use a specific set of layered products his authorisation record must have quotas to reflect that. If you remove that layered product you should adjust his quotas. Does this mean that you want to have the ability to maintain that list with each user? I think that is more general than we are looking at but we will note it. VAX-55 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session Comment from audience, unable to hear. 45. HANNU KOKKO, UNIC OY, FINLAND I would like to reuse TPU-procedures. Is there a way to do it (include call)? Include file or call or anything with TPU. PANEL MEMBER There is nobody here who is representing TPU but we will take it back. 46. TONY HUNSON, ICI. UK Will it ever be possible to close DECnet down and PSI down separately? What I want to do is close DECnet down and leave the X.25 up as long as possible. PANEL MEMBER Is there anyone from PSI in the audience? I am sorry, we can't even take that one home with us, as PSI is a European network engineering. TONY ARNOLD I have quite a nice story concerning PSI. We had a hacker get into one of our machines across X.25 and he logged in as a username called SHUTDOWN. Now if you log in as SHUTDOWN on this particular machine, you can guess what happens - the machine shutdown. Half way through this particular procedure, it switched off DECnet, switched off its PSI link and then hung, because it could no longer send its messages out to the terminal that requested the shutdown. It remained in that state for the rest of the weekend. Quite a good security feature I thought! (Laughter) Can we have another question please? BACK TO MAIN CONFERENCE VAX-56 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session (Mid question) The trouble is, with coming from Version 3 and so on we don't get users to go into and understand the requirement for first being allowed into the rights list database, access to identifiers and then they have to get the access to the files by putting another ACL on the element. Some form of niche which would allow rights list to grant access to files owned by a particular user would be very nice. PANEL MEMBER I am not sure I understand the question. If you could catch me later on that. QUESTION FORTY SIX KLAUS PETERSON, DENMARK When will system messages be available in HELP? PANEL MEMBER We will take that under advisement. I don't know whether it is now or will be in the future. KLAUS Okay, one more. What about LIB$COPY? When will that come? PANEL MEMBER There are no plans. However, making more utilities callable is a direction we are heading in. 47. ULRICH STRATHMEIER, BEB HANNON, GERMANY What do you think of replacing the system UIC by a resource identifier? PANEL MEMBER VAX-57 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session ULRICH To allow several people to do system management functions. PANEL MEMBER Group management capabilities are significant continuing enhancements to the security capabilities. This is not to be alluded to in the next future release but shortly after that you will see significant extensions to VMS security capabilities. I can't answer whether what you want specifically will be addressed. Certainly the issue of allowing group management of the resources and protection to the system is part of that work. EVERYTHING GOES QUIET, THEN LAUGHTER AND APPLAUSE. UNABLE TO ASCERTAIN WHETHER I HAVE MISSED ANY QUESTIONS OR NOT. 48. I am a very new user of VAX. I got my VAX service in the spring. The VAX serviceman said that with an 8 bit character code, I could use all the European letters. But when I installed all the products from this and tried the European letters, I found it is not possible. I will give you some examples. You say, relational database. I want to give my fields a decent understandable name but it is not possible. I have to use a cryptic abbreviation. This propagates throughout the system. Suppose I buy Rally. Rally uses the description in the database and puts this cryptical abbreviation on the screen and no one understands it. I have to change it to something readable, and there I lose a lot of time. I lose so much time that the Rally is no longer a fourth generation productivity tool. I find this very distressing and annoying. (applause) PANEL MEMBER Providing international collating sequences is something that we regard as very important and you can expect to see support for that in RMS index files, in the SORT utility. VAX-58 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I can't speak for the relational database product. You were referring to VAX Rdb? If there is somebody from database here who can speak for them, feel free to step forward. I can't speak for them. As well as better support, also in the terminals area. That is an important problem for us. The European community is a large portion of our customer base. 49. EHRENSPERGER KURT. FIDES INFORMATIK, ZURICH When I am doing stand-alone BACKUP on a VAX-11/750, sometimes the machine crashes. Why? It happens since VMS V4.5. PANEL MEMBER Are you running from a hard disk version stand-alone backup? EHRENSPERGER KURT Yes PANEL MEMBER Are you restoring to the same disk you booted from? EHRENSPERGER KURT No, it is just an image backup from the hard disk to the tape. And it crashes. PANEL MEMBER This problem is not known. Would you put a copy of either a point or two or a dump file in the stand-alone backup SYSEXE directory. If it is your SYSE directory that means CREATE [SYSE.SYSEXE]SYSDUMP.DMP and send your dump along with your SPR. EHRENSPERGER KURT VAX-59 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I am not clear how I can get a dump on stand-alone backup. PANEL MEMBER You booted stand-alone backup on a large disc? Yes. Put a dump file on the SYSEXE directory on that large disc and when stand-alone backup crashes it will write a dump file into SYSDUMP. EHRENSPERGER KURT 50. Another small question. When I use ANALYZE/DISK_STRUCTURE and /REPAIR and so on, then I get some error messages on directory file, SYSEXE, that it has some bad point. PANEL MEMBER The back link problems? That is fixed in a future release. We begin to discriminate between the original and file directory entry and the aliases and we stop making those messages. That also means that we will no longer be alternately flipping the back links from one file to another as you backup and restore and analyse and do things like that. EHRENSPERGER KURT Will it be implemented in V4.6? No. Perhaps in V5? 51. FREEK DE DRUYF, COMP. CENTRE, DELFT UNIVERSITY OF TECHNOLOGY Will there be accounting for logical links which pass through a Router node? Gathering should take place in the router node. PANEL MEMBER You can repeat that request for future features from the DECnet people if we get the link up - if we can establish a logical link to the States! VAX-60 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session TONY ARNOLD We are currently waiting for the operators to put us through. As soon as they do it will become live, and either all hell will be let loose, or we will be talking to the States. Could you please ask any further questions? 52. When I received VMS V4 I saw something named Multinational set. I tried MAIL, SET PERSONALNAME and it rejects characters from Multinational set. You cannot use Multinational character sets with SET PERSONALNAME commands in the MAIL utility. PANEL MEMBER He has used it right there. I think I have seen it in his personal name. I have seen Multinational characters in his personal name. TONY ARNOLD Is that patch available from DEC? Could it be put on a sig tape? PANEL MEMBER It is noted as a problem. We will see if it has been fixed in 4.6. 53. ALAN SILVERMAN, CERN Another security question. One of the tools we are supposed to use was CHECKSUM checking after various incidents and it provides a good idea that you provide some more tools for security. In particular, the CHECKSUM program should get some documentation so that people will know how to use it properly. It would help us if you provided us with some examples that we could use, for example, correct checksums for images after you ship them. You could provide a file in each kit which had the checksum for the images you installed which you could then install and use. CHECKSUM sounds the perfect solution and it would be nice if you would help us to check our own system. VAX-61 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER CHECKSUMs worked this time but the algorithm is known and the next time the hackers come, if there happens to be another break-in like this, they will know enough to fix the image so that the checksum matches what it is supposed to be. If we use an encryption algorithm for the checksum, then we have a massive redistribution problem with thousands of keys and thousands of customers, and we don't know how to maintain that. ALAN SILVERMAN I agree it is not a perfect solution, but it is better than nothing. You offer nothing right now. PANEL MEMBER I think a private system where you calculated your own CHECKSUM would be more useful. If you have this need, a private image to calculate CHECKSUMs for yourself is a reasonably simple procedure. One thing I would caution; CHECKSUM is basically a component of VMS INSTALL and could change as the needs of VMS INSTALL change, so depending on the current operation of CHECKSUM might be a problem for you. QUESTION FIFTY THREE The ERASE-on-DELETE file attribute is not propagated across successive versions of the same file. Is this a security feature, or is it a bug? PANEL MEMBER I am sorry, this is not a problem with the COPY utility. You are just expecting that that bit is propagated when you create a new version of a file, like ACCESS and others are propagated. I would ask that if we can get the link up if that is part of what we consider a security attribute. That could fall either way. As a file attribute, we don't propagate file attributes, just the security attributes and I am not sure what our security people would say to that. VAX-62 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER To us non-security types it sounds reasonable so we will take that back. 54. In a large system it becomes very difficult to manipulate the device control library. We have forty or fifty printers running and it is almost never possible to manipulate the library. Do you have any solution? PANEL MEMBER I would suggest some kind of mechanism whereby a new device control library could be built and then in some easy way propagated over to the printing system. QUESTIONER At the moment you have to stop and delete all the 40 printer queues and restart the whole thing again. (An attempt to make the link up follows. Difficulty for either side to hear. Tony Arnold asked for a question.) TAPE BREAK QUESTION FIFTY FOUR I have been told of ways that the AUTOGEN procedure will change in future releases. I want to know if there is some provision to make the parameters depend on the power of the CPU. It seems there are some parameters that have different values from the standard one, depending on the power of the CPU. PANEL MEMBER Aside from QUANTUM, what else do you think would depend on the power of CPU? VAX-63 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER The reason I ask is because there are so few of them that we cannot decide ourselves on whether QUANTUM is too big or too small right now and so we have elected to leave it as it is. Because it is only one parameter. AUTOGEN would be useful if there were a large number of system parameters to change, but if you only have to change one parameter maybe there are better things to do with AUTOGEN. PANEL MEMBER The rate parameters, page fault rate and so on, are based on page faults per CPU second so they naturally adjust. If CPU runs faster you get more page faults and faster adjustment. QUESTIONER My second question is about wild card characters. I mean, for example, characters to exclude specific beginnings for files or something like that. PANEL MEMBER Noted as a suggestion. There are no plans at present to provide an extension to the wild card capabilities in file names. QUESTIONER Another question is something like an INCLUDE command to be able to expand a file dynamically. It is possible of course in the DCL level with an @ command, but anywhere, even if you are executing a user name. PANEL MEMBER There is an RTL routine to do wild card file processing, LIB$FILESCAN, but you are talking about interactively at the terminal? Noted. VAX-64 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER A comment on the last two. Those and some of the others in the last half hour or so have been more on the suggestion type side than question side, so I would add to that, submit an SIR. TONY ARNOLD Yes, if you could please gear your questions to problems that are of a technical nature and submit an SIR with suggestions. 55. When you look into VAX VMS system, the password is staying in the typeahead for a while. This is not good security and should be purged automatically by a login autotext. Will there be something done in a future release? You can certainly write an application doing IO$PURGE but it should be done automatically. PANEL MEMBER On the surface that would seem like a very straightforward problem to solve. However, the terminal typeahead buffer is only readable from inner modes, and people with enough privilege to read the type ahead buffer have enough privilege to damage you in so many ways that closing that one is insignificant. QUESTIONER I am thinking that when the system crashes on mag tape, the password remains on that tape. PANEL MEMBER Oh we have lots more information than passwords when you send us a crash dump! There are all kinds of secure information available all through a crash dump that is now completely at the disposal of the reader. That is just one of many problems. When you release the crash dump you are releasing a lot of information. Please send us more dumps! (laughter) VAX-65 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER And your modem phone numbers! 56. I want to ask a question about the Fortran compiler. Presently we do not have different levels of conversation, only talking nice or not talking nice. Is there going to be anything done about it? And will there be a qualifier to expand in line code of codes to library logins? PANEL MEMBER Were those both Fortran questions? Anyone from VAX languages here that can comment on that? PANEL MEMBER Jim Totten is the person that you need to speak to. He is the person from the Languages Group. He will be in the exhibition tomorrow. 57. Using SET TERMINAL/INQUIRE on the window on the VAX station, assuming the window has a size of more than 24 lines, SET TERMINAL/INQUIRE resets 24 lines for that window. Are there any plans to change that behaviour? I think it is bad behaviour. PANEL MEMBER I think that is bad behaviour too. We will take that back. PANEL MEMBER A question though. What good is it? QUESTIONER VAX-66 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session I think it is good to... PANEL MEMBER For a workstation window, usually all of the attributes are already known, so enquiring what they are is... QUESTIONER Supposing you are already logged in and you set host to another node. You login and you have SET TERMINAL/INQUIRE, then you change the size of your window. PANEL MEMBER Fix this your LOGIN.COM. We solved this problem by having our LOGIN.COMs bypass a SET TERM/INQUIRE for a remote terminal. My workstation never does a SET TERM/ INQUIRE because it doesn't do anything useful. QUESTION FIFTY SEVEN PAUL TEECE, GOLDMAN SACHS, U.K. This is a quick question, probably as much to the floor actually. Does anyone know of a VAX version of VI, the UNIX editor. DEC says it may be the version of DECSHELL under VMS V5. But it will be neater if there is a TPU script available. TONY ARNOLD Can I ask a stupid question? Who on earth wants to use VI? PAUL TEECE We have some recidivist UNIX guys who will only go VMS if VI is available. TONY ARNOLD Teach them EMACS and put EMACS on the VMS. VAX-67 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PAUL TEECE We have EMACS, we had to get that as well. TONY ARNOLD This is a good question. Can I generalise the question a little bit and ask what plans there are for other interfaces for TPU? We have EVE, we have EDT, WPS+ has been mentioned. Are there any plans for things like an EMACS emulator, or a VI emulator... or is there an EVE emulator in TECO? PANEL MEMBER I don't think you have anyone here who can comment on TPU features. We can say there are no new files in Version Future. 58. BURT ZORN, T.U. DELFT, NETHERLANDS Why can't virtual terminals be enabled on RTA terminals? TONY ARNOLD Disconnectable? You mean so that you can disconnect? BURT Yes. PANEL MEMBER It can't be done, and he doesn't know why. TONY ARNOLD I think I know somebody who has done it. BURT I have already tried in the SYSGEN parameters but they don't seem to work in the RT terminals. VAX-68 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER He remembered why after he told him. The class driver which implements virtual terminals is only running on the original host, not on the host that you SET HOST to. That is why we haven't done it. If someone has managed to seriously subvert the will and intent of a remote terminal.... TONY ARNOLD I think the way it was done was to set the bit in the default terminal characteristics...that doesn't work. PANEL MEMBER It works for LAT terminals. TONY ARNOLD It may have been an X.29 call that he did it for, over X.25. PANEL MEMBER I think that that is on the "it can't be done without a major redesign" list. No such redesign is planned. 59. PAUL TEECE, GOLDMAN SACHS, U.K. Since the US link failed could you very quickly say whether you think there are any changes to the DEBNT driver in 4.6? PANEL MEMBER There are changes to ET driver in 4.6. What question were you looking for? PAUL TEECE My question was about using TSM and NCP to connect to terminal servers and the problem is to do with the DEBNT driver, we believe. If you are saying that there are changes in 4.6... VAX-69 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session PANEL MEMBER There are a large number of Ethernet driver changes in 4.6. These are primarily changes prompted by the excessive stress that the introduction of LAVC made to those drivers. So the various 4.5 local area VAX cluster Ethernet drivers were included in 4.6. 60. BANCK, PER MAGNUS, GREATER STOCKHOLM TRANSPORT A captive user is, through legitimate use of EDT on a specified file, able to INCLUDE and WRITE another file. Is it possible to specify a nowrite bit on calling EDT and other editors? PANEL MEMBER We have one similar problem in an application that we use for our field test problem reporting database. The technique I used there was to have the problem database look only at Version 1 of significant files, so that the user putting random garbage out there is ignored by the applications running. We can note that as a problem. Do you want to say, prevent a user from using EDT, mail? PER MAGNUS No, really, to prevent a user from using the WRITE statement within EDT. To suppress that use because that is what circumvents his captive state. I want him to write out the file he is editing or whatever, but I don't want him to go into additional files and then be able to write them back. PANEL MEMBER That would imply nowrite and no specification of a file name on the EXIT. PER MAGNUS EXIT name is all right, because that is what he is in, for whatever reason, but what he can do as well is to just do INCLUDE, for example a login file and then write it back. Then he can EXIT the ordinary way and no one knows how he was able to change the login file. VAX-70 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session These are captive accounts I am aware of. PANEL MEMBER I am assuming the same problem must exist with TPU or almost any other editor. It is sort of a generic problem in that you are in a captive account, allowing someone to get at something which is not very captive. PANEL MEMBER I would say if you have a captive account whose login command file is LOGIN.COM and it is in a writable directory, that might be your problem. We don't have any LOGIN.COMs accessible from EDIT, and the captive command file is in a protected directory. AUDIENCE MEMBER We solved that problem by making the default directory owned by somebody else, so that the captive account actually didn't have permission to write into it at all. TONY ARNOLD Could you also set your LOGIN.COM to be execute only? I believe that works. 61. If a file is locked by another user it is impossible to copy or type or something like that, but it is possible to RENAME. Why is this possible? PANEL MEMBER Your question is, why is a user allowed to RENAME a file that is protected? QUESTIONER VAX-71 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session Locked by another user - opened by another user or something like that. PANEL MEMBER Sounds like a feature to me! (laughter) You can RENAME it with a simple XQP call which is what the RENAME ends up doing. It does not require you to open the file for reading. That is where you get your access conflict, when you try read and somebody else has the file open for exclusive operation. You don't have to open the file to do in order to RENAME it. QUESTIONER Yes, but I expect when I try to RENAME a file that is locked by another user, it shouldn't be RENAMEd. PANEL MEMBER There is no problem at all because the name of a file is not bound to the access to the file. That is all handled by the file ID and once that is allocated, how the name is bound to the ID is a separate operation of entering stuff in the directories and can proceed independently. PANEL MEMBER As an aside on that, in Version Future, the ability to manipulate the filename is controlled by the protection of the file. That is a change where currently you can RENAME a file but you cannot delete, and effectively delete it because you remove the name. We are no longer going to allow that. QUESTION SIXTY ONE PETER BIDDEN, EUROPEAN MOLECULAR BIOLOGY A little MAIL question! The scenario is quite simple. You have a four storey building. You have people on each floor. On each floor there is a printer. Now the man on the top floor would like his MAIL/PRINT to come out on his printer and the man on the ground floor would like it to come out on the ground floor printer. I asked the Telephone Service Centre in Munich and they VAX-72 PAGESWAPPER - January 1988 - Volume 9 Number 6 DECUS European VAX SIG Question and Answer Session said, "Ah yes, there is a MAIL log or something, we'll call you back." And after a day or so I called them back and they said "Whoops!" Why? Is this something that is in Version Future that hasn't been released yet? PANEL MEMBER I believe that in Version Future you will see a number of qualifiers on that or a selection ability for mail printing, specifying queue names, for example. I am told that is also likely to be part of the user profile, so you will be able to set it up on a per-user basis. TONY ARNOLD Are there any more questions? Okay, if there are no more questions, thank you very much and thank you to the panel for attending. VAX-73 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results Fall 1987 US VAX Systems SIG SIR Ballot Results Mark Oakley SIR Coordinator A total of 323 ballots were counted in the latest SIR BALLOT, a decline from the past two ballots. There was an increase in the number of ballots cast electronically on the Pageswapper Vax. Many thanks to Larry Kilgallen for the use of this machine and for his "ballot-casting" procedure. There was a tie in the voting for tenth place between SIR F87-14 (provide simple project accounting) and SIR F87-54 (send more developers to symposia). I have asked Digital to respond to both SIR's. I am hopeful that participation will continue to build and will look for ways to encourage you to vote. The SIR ballot is the only on-going program by which the SIG provides input to Digital. Top 10 (and other) SIR's continue to be incorporated into VMS. Digital has repeatedly encouraged the use of this channel of communication. The summary of this voting appears below. Digital's response to the top 11 requests overall will be presented at the Fall 1987 DECUS Symposium in Anaheim. Interpreting the SIR Ballot Results The results of the System Improvement Request ballot are shown on the following pages. All of the reports have the same one page format. Following the report title is the number of ballots counted for that report. The number shown on the "All Users" report is the total number of ballots which were returned. The ballots on the "11/780 Users" report is the number of ballots which checked the "11/780" blank on the ballot questionnaire, and so on. The SIR's are listed on the page in order of points received, from highest to lowest. The entry for each SIR begins with the SIR number (from the ballot), a brief description, and the total number of votes (positive and negative) received by that SIR. VAX-74 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results The data is summarized in two different ways. First, there are a series of reports broken down by user category. The user categories are defined by the questionnaire portion of the SIR ballot. A ballot was counted in each user category which was checked off, for example "11/780 user". Finally, there are a series of reports ranking the SIR's within SIR category. The SIR categories are those shown on the ballot, for example "DCL and Utilities" and "Commercial". The reports by SIR category use the data from all ballots received. VAX-75 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY ALL USERS Total ballots in this category: 323 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 173 24 Capture interactive session better 146 67 Individual field reports from AUTHORIZE 124 13 Standardize format for printable output 97 15 First/last file-specs in BACKUP log 92 26 Various VMSmail enhancements 91 11 Queued requests for ALLOCATE command 89 71 Provide more LAT accounting info 89 2 Implement cluster-accessible tape drives 83 14 Support for simple project accounting 78 54 Send more developers to Symposia 78 12 Tape automatic volume recognition 76 16 Support no CR/LF in DCL WRITE 76 19 Improve DCLTABLES manipulations 73 6 More detail in VMS accounting records 71 20 DCL return status enhancements 71 25 Enhance sysgen parameter readability 67 5 Abbreviated commands in DTR 64 42 Implement SYSGEN DISCONNECT 63 8 Improve VMS Mount service messages 62 62 Security alarm messages to a file. 61 64 Lexical for getting RIGHTSLIST info 59 23 DCL /LOG needs more consistency 57 34 Combine DEFINE, ASSIGN commands 57 55 Provide sales update for customers 56 58 Enhance COPY to copy ACL's 55 59 Enhance file access from user images 55 27 Enhance SET/HOST/DTE for more modems 55 35 Provide F$FILE_ATTRIBUTES service 54 37 $TRNLNM wildcard support 53 1 Improve heterogeneous user group support 52 18 Enhance command line editing 52 63 DECnet proxy access for SET HOST command 51 52 TPU should display all characters 51 30 DEFINE/KEY enhancements 47 VAX-76 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY VAX 8700/8800 USERS Total ballots in this category: 44 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 20 71 Provide more LAT accounting info 19 12 Tape automatic volume recognition 18 24 Capture interactive session better 18 6 More detail in VMS accounting records 17 67 Individual field reports from AUTHORIZE 17 11 Queued requests for ALLOCATE command 17 14 Support for simple project accounting 15 15 First/last file-specs in BACKUP log 15 10 Provide callable OPCOM interface 14 54 Send more developers to Symposia 14 64 Lexical for getting RIGHTSLIST info 13 59 Enhance file access from user images 12 5 Abbreviated commands in DTR 12 62 Security alarm messages to a file. 11 2 Implement cluster-accessible tape drives 11 1 Improve heterogeneous user group support 10 61 End-to-end encryption within DECnet-VAX 9 18 Enhance command line editing 9 29 Enhance MOUNT/FOR for uninitialized tapes 9 8 Improve VMS Mount service messages 9 26 Various VMSmail enhancements 9 20 DCL return status enhancements 8 13 Standardize format for printable output 8 69 Enhance MONITOR DISK 8 19 Improve DCLTABLES manipulations 8 55 Provide sales update for customers 7 37 $TRNLNM wildcard support 7 68 Provide HSC tape-to-tape utility 7 52 TPU should display all characters 7 4 Provide SCS communication services 7 30 DEFINE/KEY enhancements 6 35 Provide F$FILE_ATTRIBUTES service 6 23 DCL /LOG needs more consistency 6 63 DECnet proxy access for SET HOST command 6 VAX-77 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY VAX 8600/8650 USERS Total ballots in this category: 102 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 52 6 More detail in VMS accounting records 39 12 Tape automatic volume recognition 37 71 Provide more LAT accounting info 37 67 Individual field reports from AUTHORIZE 36 24 Capture interactive session better 36 11 Queued requests for ALLOCATE command 34 62 Security alarm messages to a file. 32 13 Standardize format for printable output 30 19 Improve DCLTABLES manipulations 29 2 Implement cluster-accessible tape drives 27 54 Send more developers to Symposia 25 8 Improve VMS Mount service messages 25 10 Provide callable OPCOM interface 24 64 Lexical for getting RIGHTSLIST info 24 4 Provide SCS communication services 23 15 First/last file-specs in BACKUP log 23 42 Implement SYSGEN DISCONNECT 22 26 Various VMSmail enhancements 22 14 Support for simple project accounting 22 16 Support no CR/LF in DCL WRITE 21 69 Enhance MONITOR DISK 21 59 Enhance file access from user images 21 18 Enhance command line editing 20 25 Enhance sysgen parameter readability 20 58 Enhance COPY to copy ACL's 20 52 TPU should display all characters 19 68 Provide HSC tape-to-tape utility 18 61 End-to-end encryption within DECnet-VAX 18 20 DCL return status enhancements 18 1 Improve heterogeneous user group support 17 5 Abbreviated commands in DTR 17 35 Provide F$FILE_ATTRIBUTES service 17 37 $TRNLNM wildcard support 17 63 DECnet proxy access for SET HOST command 16 VAX-78 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY VAX 8500/8550 USERS Total ballots in this category: 38 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 71 Provide more LAT accounting info 17 24 Capture interactive session better 14 28 Display open files on SHOW PROCESS 14 12 Tape automatic volume recognition 14 54 Send more developers to Symposia 12 67 Individual field reports from AUTHORIZE 12 11 Queued requests for ALLOCATE command 12 55 Provide sales update for customers 11 58 Enhance COPY to copy ACL's 11 19 Improve DCLTABLES manipulations 11 4 Provide SCS communication services 11 13 Standardize format for printable output 10 16 Support no CR/LF in DCL WRITE 10 1 Improve heterogeneous user group support 9 27 Enhance SET/HOST/DTE for more modems 9 2 Implement cluster-accessible tape drives 9 63 DECnet proxy access for SET HOST command 9 34 Combine DEFINE, ASSIGN commands 9 52 TPU should display all characters 9 7 Better terminal-printer support 8 6 More detail in VMS accounting records 8 62 Security alarm messages to a file. 8 15 First/last file-specs in BACKUP log 7 26 Various VMSmail enhancements 7 8 Improve VMS Mount service messages 7 5 Abbreviated commands in DTR 7 30 DEFINE/KEY enhancements 7 20 DCL return status enhancements 7 68 Provide HSC tape-to-tape utility 7 23 DCL /LOG needs more consistency 7 61 End-to-end encryption within DECnet-VAX 6 41 Support for higher baud rates 6 35 Provide F$FILE_ATTRIBUTES service 5 22 /USER for running detached jobs 5 31 Add /BELL to various DCL commands 5 VAX-79 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY VAX 8300/8200 USERS Total ballots in this category: 61 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 24 Capture interactive session better 31 28 Display open files on SHOW PROCESS 27 71 Provide more LAT accounting info 27 15 First/last file-specs in BACKUP log 22 26 Various VMSmail enhancements 21 67 Individual field reports from AUTHORIZE 20 54 Send more developers to Symposia 18 6 More detail in VMS accounting records 16 2 Implement cluster-accessible tape drives 16 11 Queued requests for ALLOCATE command 15 59 Enhance file access from user images 15 12 Tape automatic volume recognition 15 1 Improve heterogeneous user group support 15 19 Improve DCLTABLES manipulations 14 4 Provide SCS communication services 14 13 Standardize format for printable output 14 5 Abbreviated commands in DTR 14 27 Enhance SET/HOST/DTE for more modems 13 68 Provide HSC tape-to-tape utility 13 25 Enhance sysgen parameter readability 13 55 Provide sales update for customers 12 8 Improve VMS Mount service messages 12 62 Security alarm messages to a file. 12 37 $TRNLNM wildcard support 12 42 Implement SYSGEN DISCONNECT 12 14 Support for simple project accounting 12 56 Enhance DECnet remote file access 11 10 Provide callable OPCOM interface 11 16 Support no CR/LF in DCL WRITE 11 18 Enhance command line editing 10 61 End-to-end encryption within DECnet-VAX 10 52 TPU should display all characters 10 63 DECnet proxy access for SET HOST command 10 64 Lexical for getting RIGHTSLIST info 10 66 STB support on more devices 10 VAX-80 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY 11/780, 11/782 and 11/785 USER Total ballots in this category: 198 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 106 24 Capture interactive session better 77 67 Individual field reports from AUTHORIZE 76 11 Queued requests for ALLOCATE command 68 2 Implement cluster-accessible tape drives 67 13 Standardize format for printable output 58 15 First/last file-specs in BACKUP log 56 12 Tape automatic volume recognition 56 71 Provide more LAT accounting info 54 54 Send more developers to Symposia 51 26 Various VMSmail enhancements 50 6 More detail in VMS accounting records 49 19 Improve DCLTABLES manipulations 47 16 Support no CR/LF in DCL WRITE 45 5 Abbreviated commands in DTR 45 62 Security alarm messages to a file. 44 8 Improve VMS Mount service messages 42 14 Support for simple project accounting 42 20 DCL return status enhancements 41 25 Enhance sysgen parameter readability 41 55 Provide sales update for customers 39 42 Implement SYSGEN DISCONNECT 37 1 Improve heterogeneous user group support 37 64 Lexical for getting RIGHTSLIST info 36 59 Enhance file access from user images 36 18 Enhance command line editing 36 34 Combine DEFINE, ASSIGN commands 34 10 Provide callable OPCOM interface 33 30 DEFINE/KEY enhancements 32 63 DECnet proxy access for SET HOST command 31 52 TPU should display all characters 31 61 End-to-end encryption within DECnet-VAX 30 23 DCL /LOG needs more consistency 30 56 Enhance DECnet remote file access 29 68 Provide HSC tape-to-tape utility 29 VAX-81 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY 11/750 USERS Total ballots in this category: 158 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 82 24 Capture interactive session better 63 67 Individual field reports from AUTHORIZE 54 54 Send more developers to Symposia 49 71 Provide more LAT accounting info 49 14 Support for simple project accounting 44 11 Queued requests for ALLOCATE command 44 13 Standardize format for printable output 43 26 Various VMSmail enhancements 43 15 First/last file-specs in BACKUP log 43 2 Implement cluster-accessible tape drives 39 16 Support no CR/LF in DCL WRITE 39 25 Enhance sysgen parameter readability 35 1 Improve heterogeneous user group support 34 34 Combine DEFINE, ASSIGN commands 34 42 Implement SYSGEN DISCONNECT 34 20 DCL return status enhancements 33 27 Enhance SET/HOST/DTE for more modems 33 19 Improve DCLTABLES manipulations 32 5 Abbreviated commands in DTR 31 58 Enhance COPY to copy ACL's 31 35 Provide F$FILE_ATTRIBUTES service 31 6 More detail in VMS accounting records 31 63 DECnet proxy access for SET HOST command 30 30 DEFINE/KEY enhancements 29 52 TPU should display all characters 28 64 Lexical for getting RIGHTSLIST info 28 66 STB support on more devices 28 23 DCL /LOG needs more consistency 28 12 Tape automatic volume recognition 28 59 Enhance file access from user images 25 55 Provide sales update for customers 25 37 $TRNLNM wildcard support 25 22 /USER for running detached jobs 22 3 Remove TU-58 boot requirement for 750 22 VAX-82 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY 11/730 and 11/725 USERS Total ballots in this category: 69 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 24 Capture interactive session better 37 28 Display open files on SHOW PROCESS 37 67 Individual field reports from AUTHORIZE 25 20 DCL return status enhancements 22 13 Standardize format for printable output 22 11 Queued requests for ALLOCATE command 21 26 Various VMSmail enhancements 20 5 Abbreviated commands in DTR 18 71 Provide more LAT accounting info 18 14 Support for simple project accounting 16 54 Send more developers to Symposia 16 2 Implement cluster-accessible tape drives 15 16 Support no CR/LF in DCL WRITE 15 35 Provide F$FILE_ATTRIBUTES service 15 12 Tape automatic volume recognition 15 55 Provide sales update for customers 15 62 Security alarm messages to a file. 15 6 More detail in VMS accounting records 15 25 Enhance sysgen parameter readability 15 64 Lexical for getting RIGHTSLIST info 14 1 Improve heterogeneous user group support 14 15 First/last file-specs in BACKUP log 14 7 Better terminal-printer support 13 27 Enhance SET/HOST/DTE for more modems 13 37 $TRNLNM wildcard support 13 52 TPU should display all characters 12 59 Enhance file access from user images 12 19 Improve DCLTABLES manipulations 12 22 /USER for running detached jobs 11 8 Improve VMS Mount service messages 11 56 Enhance DECnet remote file access 11 69 Enhance MONITOR DISK 11 10 Provide callable OPCOM interface 11 61 End-to-end encryption within DECnet-VAX 10 30 DEFINE/KEY enhancements 10 VAX-83 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY MicroVAX I,II USERS Total ballots in this category: 200 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 102 24 Capture interactive session better 95 67 Individual field reports from AUTHORIZE 84 71 Provide more LAT accounting info 58 11 Queued requests for ALLOCATE command 57 13 Standardize format for printable output 57 54 Send more developers to Symposia 55 26 Various VMSmail enhancements 54 16 Support no CR/LF in DCL WRITE 53 15 First/last file-specs in BACKUP log 52 2 Implement cluster-accessible tape drives 52 25 Enhance sysgen parameter readability 49 14 Support for simple project accounting 48 37 $TRNLNM wildcard support 46 35 Provide F$FILE_ATTRIBUTES service 46 19 Improve DCLTABLES manipulations 44 6 More detail in VMS accounting records 44 8 Improve VMS Mount service messages 43 20 DCL return status enhancements 42 12 Tape automatic volume recognition 42 42 Implement SYSGEN DISCONNECT 41 55 Provide sales update for customers 41 63 DECnet proxy access for SET HOST command 40 64 Lexical for getting RIGHTSLIST info 40 62 Security alarm messages to a file. 38 5 Abbreviated commands in DTR 35 27 Enhance SET/HOST/DTE for more modems 34 18 Enhance command line editing 33 58 Enhance COPY to copy ACL's 32 1 Improve heterogeneous user group support 32 52 TPU should display all characters 31 23 DCL /LOG needs more consistency 31 47 Resolve Fortran/Debug "TRUE" 31 4 Provide SCS communication services 30 59 Enhance file access from user images 29 VAX-84 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY MicroVAX 2000, VAXstation 2000 Total ballots in this category: 45 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 24 Capture interactive session better 17 28 Display open files on SHOW PROCESS 17 2 Implement cluster-accessible tape drives 16 71 Provide more LAT accounting info 16 67 Individual field reports from AUTHORIZE 15 54 Send more developers to Symposia 14 18 Enhance command line editing 13 35 Provide F$FILE_ATTRIBUTES service 12 5 Abbreviated commands in DTR 12 55 Provide sales update for customers 12 59 Enhance file access from user images 12 27 Enhance SET/HOST/DTE for more modems 12 1 Improve heterogeneous user group support 12 61 End-to-end encryption within DECnet-VAX 11 15 First/last file-specs in BACKUP log 11 25 Enhance sysgen parameter readability 11 14 Support for simple project accounting 10 62 Security alarm messages to a file. 10 64 Lexical for getting RIGHTSLIST info 10 12 Tape automatic volume recognition 10 16 Support no CR/LF in DCL WRITE 10 4 Provide SCS communication services 9 58 Enhance COPY to copy ACL's 9 7 Better terminal-printer support 9 19 Improve DCLTABLES manipulations 9 22 /USER for running detached jobs 9 37 $TRNLNM wildcard support 9 52 TPU should display all characters 9 11 Queued requests for ALLOCATE command 9 26 Various VMSmail enhancements 8 10 Provide callable OPCOM interface 8 13 Standardize format for printable output 8 34 Combine DEFINE, ASSIGN commands 7 8 Improve VMS Mount service messages 7 56 Enhance DECnet remote file access 7 VAX-85 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY BUSINESS EDP USERS Total ballots in this category: 103 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 62 24 Capture interactive session better 52 67 Individual field reports from AUTHORIZE 40 13 Standardize format for printable output 33 5 Abbreviated commands in DTR 33 15 First/last file-specs in BACKUP log 32 20 DCL return status enhancements 30 71 Provide more LAT accounting info 30 2 Implement cluster-accessible tape drives 28 12 Tape automatic volume recognition 28 26 Various VMSmail enhancements 28 16 Support no CR/LF in DCL WRITE 27 8 Improve VMS Mount service messages 25 11 Queued requests for ALLOCATE command 25 54 Send more developers to Symposia 24 6 More detail in VMS accounting records 24 25 Enhance sysgen parameter readability 24 19 Improve DCLTABLES manipulations 21 64 Lexical for getting RIGHTSLIST info 20 23 DCL /LOG needs more consistency 20 63 DECnet proxy access for SET HOST command 20 10 Provide callable OPCOM interface 19 62 Security alarm messages to a file. 19 58 Enhance COPY to copy ACL's 18 69 Enhance MONITOR DISK 18 29 Enhance MOUNT/FOR for uninitialized tapes 18 1 Improve heterogeneous user group support 16 34 Combine DEFINE, ASSIGN commands 16 42 Implement SYSGEN DISCONNECT 16 50 Provide traceback on PERFORM statements 16 9 Provide a fast file scan. 16 7 Better terminal-printer support 16 55 Provide sales update for customers 14 68 Provide HSC tape-to-tape utility 14 56 Enhance DECnet remote file access 14 VAX-86 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY SOFTWARE DEVELOPERS Total ballots in this category: 246 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 132 24 Capture interactive session better 107 67 Individual field reports from AUTHORIZE 86 11 Queued requests for ALLOCATE command 71 2 Implement cluster-accessible tape drives 71 13 Standardize format for printable output 67 26 Various VMSmail enhancements 64 71 Provide more LAT accounting info 63 16 Support no CR/LF in DCL WRITE 62 12 Tape automatic volume recognition 61 14 Support for simple project accounting 61 54 Send more developers to Symposia 60 15 First/last file-specs in BACKUP log 59 20 DCL return status enhancements 56 19 Improve DCLTABLES manipulations 55 55 Provide sales update for customers 52 62 Security alarm messages to a file. 51 25 Enhance sysgen parameter readability 50 5 Abbreviated commands in DTR 50 6 More detail in VMS accounting records 47 64 Lexical for getting RIGHTSLIST info 47 35 Provide F$FILE_ATTRIBUTES service 46 37 $TRNLNM wildcard support 46 42 Implement SYSGEN DISCONNECT 46 8 Improve VMS Mount service messages 45 1 Improve heterogeneous user group support 42 59 Enhance file access from user images 42 63 DECnet proxy access for SET HOST command 41 58 Enhance COPY to copy ACL's 41 30 DEFINE/KEY enhancements 41 18 Enhance command line editing 41 27 Enhance SET/HOST/DTE for more modems 39 23 DCL /LOG needs more consistency 39 52 TPU should display all characters 39 47 Resolve Fortran/Debug "TRUE" 37 VAX-87 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY EDUCATIONAL USERS Total ballots in this category: 62 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 34 24 Capture interactive session better 32 67 Individual field reports from AUTHORIZE 26 15 First/last file-specs in BACKUP log 21 6 More detail in VMS accounting records 21 16 Support no CR/LF in DCL WRITE 20 13 Standardize format for printable output 20 71 Provide more LAT accounting info 20 26 Various VMSmail enhancements 19 19 Improve DCLTABLES manipulations 18 12 Tape automatic volume recognition 18 11 Queued requests for ALLOCATE command 18 14 Support for simple project accounting 16 64 Lexical for getting RIGHTSLIST info 16 63 DECnet proxy access for SET HOST command 15 2 Implement cluster-accessible tape drives 15 29 Enhance MOUNT/FOR for uninitialized tapes 15 62 Security alarm messages to a file. 15 44 Add sorting capability in editors 14 59 Enhance file access from user images 13 54 Send more developers to Symposia 13 43 Provide "footers" in RUNOFF 12 25 Enhance sysgen parameter readability 12 5 Abbreviated commands in DTR 12 55 Provide sales update for customers 12 37 $TRNLNM wildcard support 12 10 Provide callable OPCOM interface 11 58 Enhance COPY to copy ACL's 11 34 Combine DEFINE, ASSIGN commands 11 1 Improve heterogeneous user group support 11 20 DCL return status enhancements 10 23 DCL /LOG needs more consistency 10 18 Enhance command line editing 10 31 Add /BELL to various DCL commands 10 8 Improve VMS Mount service messages 10 VAX-88 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY COMPUTER SCI. RESEARCHERS Total ballots in this category: 30 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 12 Tape automatic volume recognition 15 11 Queued requests for ALLOCATE command 14 6 More detail in VMS accounting records 12 54 Send more developers to Symposia 12 19 Improve DCLTABLES manipulations 11 71 Provide more LAT accounting info 11 24 Capture interactive session better 10 67 Individual field reports from AUTHORIZE 10 28 Display open files on SHOW PROCESS 10 14 Support for simple project accounting 9 15 First/last file-specs in BACKUP log 8 34 Combine DEFINE, ASSIGN commands 8 16 Support no CR/LF in DCL WRITE 8 55 Provide sales update for customers 8 1 Improve heterogeneous user group support 8 5 Abbreviated commands in DTR 8 29 Enhance MOUNT/FOR for uninitialized tapes 7 58 Enhance COPY to copy ACL's 7 27 Enhance SET/HOST/DTE for more modems 7 13 Standardize format for printable output 7 63 DECnet proxy access for SET HOST command 6 10 Provide callable OPCOM interface 6 2 Implement cluster-accessible tape drives 6 56 Enhance DECnet remote file access 5 30 DEFINE/KEY enhancements 5 60 Mandatory security controls in VMS. 5 61 End-to-end encryption within DECnet-VAX 5 8 Improve VMS Mount service messages 5 64 Lexical for getting RIGHTSLIST info 5 66 STB support on more devices 5 4 Provide SCS communication services 5 26 Various VMSmail enhancements 5 22 /USER for running detached jobs 4 37 $TRNLNM wildcard support 4 41 Support for higher baud rates 4 VAX-89 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY DATA ACQ./CTRL USERS Total ballots in this category: 64 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 31 54 Send more developers to Symposia 26 13 Standardize format for printable output 22 71 Provide more LAT accounting info 21 67 Individual field reports from AUTHORIZE 20 19 Improve DCLTABLES manipulations 19 24 Capture interactive session better 19 20 DCL return status enhancements 18 58 Enhance COPY to copy ACL's 17 59 Enhance file access from user images 17 41 Support for higher baud rates 17 14 Support for simple project accounting 17 27 Enhance SET/HOST/DTE for more modems 16 1 Improve heterogeneous user group support 16 35 Provide F$FILE_ATTRIBUTES service 16 30 DEFINE/KEY enhancements 15 11 Queued requests for ALLOCATE command 14 66 STB support on more devices 14 55 Provide sales update for customers 14 47 Resolve Fortran/Debug "TRUE" 14 37 $TRNLNM wildcard support 13 26 Various VMSmail enhancements 13 8 Improve VMS Mount service messages 12 25 Enhance sysgen parameter readability 12 16 Support no CR/LF in DCL WRITE 12 2 Implement cluster-accessible tape drives 12 63 DECnet proxy access for SET HOST command 12 64 Lexical for getting RIGHTSLIST info 12 45 Routine to interpret IOSTAT 12 6 More detail in VMS accounting records 12 52 TPU should display all characters 12 46 Standardize data-type support 10 61 End-to-end encryption within DECnet-VAX 10 62 Security alarm messages to a file. 10 34 Combine DEFINE, ASSIGN commands 10 VAX-90 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY CAD/CAM USERS Total ballots in this category: 72 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 67 Individual field reports from AUTHORIZE 40 24 Capture interactive session better 34 28 Display open files on SHOW PROCESS 31 11 Queued requests for ALLOCATE command 28 71 Provide more LAT accounting info 25 12 Tape automatic volume recognition 22 14 Support for simple project accounting 22 13 Standardize format for printable output 21 15 First/last file-specs in BACKUP log 21 6 More detail in VMS accounting records 20 2 Implement cluster-accessible tape drives 20 26 Various VMSmail enhancements 18 54 Send more developers to Symposia 18 16 Support no CR/LF in DCL WRITE 17 35 Provide F$FILE_ATTRIBUTES service 16 5 Abbreviated commands in DTR 16 59 Enhance file access from user images 16 61 End-to-end encryption within DECnet-VAX 16 10 Provide callable OPCOM interface 16 19 Improve DCLTABLES manipulations 16 55 Provide sales update for customers 15 20 DCL return status enhancements 14 1 Improve heterogeneous user group support 14 64 Lexical for getting RIGHTSLIST info 14 25 Enhance sysgen parameter readability 14 8 Improve VMS Mount service messages 14 30 DEFINE/KEY enhancements 13 42 Implement SYSGEN DISCONNECT 11 52 TPU should display all characters 11 18 Enhance command line editing 11 41 Support for higher baud rates 11 70 More attributes for BACKUP output files 11 56 Enhance DECnet remote file access 11 27 Enhance SET/HOST/DTE for more modems 10 23 DCL /LOG needs more consistency 9 VAX-91 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY SERVICE BUREAU OPERATORS Total ballots in this category: 26 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 16 12 Tape automatic volume recognition 11 2 Implement cluster-accessible tape drives 9 9 Provide a fast file scan. 9 5 Abbreviated commands in DTR 8 20 DCL return status enhancements 8 24 Capture interactive session better 8 1 Improve heterogeneous user group support 8 55 Provide sales update for customers 8 67 Individual field reports from AUTHORIZE 8 6 More detail in VMS accounting records 7 4 Provide SCS communication services 7 50 Provide traceback on PERFORM statements 6 19 Improve DCLTABLES manipulations 6 8 Improve VMS Mount service messages 6 68 Provide HSC tape-to-tape utility 6 11 Queued requests for ALLOCATE command 5 54 Send more developers to Symposia 5 26 Various VMSmail enhancements 5 62 Security alarm messages to a file. 5 22 /USER for running detached jobs 5 46 Standardize data-type support 5 71 Provide more LAT accounting info 5 58 Enhance COPY to copy ACL's 4 25 Enhance sysgen parameter readability 4 23 DCL /LOG needs more consistency 4 16 Support no CR/LF in DCL WRITE 4 29 Enhance MOUNT/FOR for uninitialized tapes 4 15 First/last file-specs in BACKUP log 3 27 Enhance SET/HOST/DTE for more modems 3 7 Better terminal-printer support 3 13 Standardize format for printable output 3 61 End-to-end encryption within DECnet-VAX 3 30 DEFINE/KEY enhancements 3 35 Provide F$FILE_ATTRIBUTES service 3 VAX-92 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY HARDWARE DEVELOPERS Total ballots in this category: 35 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 11 Queued requests for ALLOCATE command 15 24 Capture interactive session better 15 28 Display open files on SHOW PROCESS 15 67 Individual field reports from AUTHORIZE 14 54 Send more developers to Symposia 13 2 Implement cluster-accessible tape drives 13 26 Various VMSmail enhancements 11 12 Tape automatic volume recognition 11 55 Provide sales update for customers 10 62 Security alarm messages to a file. 10 65 Identifier to be the owner of a process 10 6 More detail in VMS accounting records 10 61 End-to-end encryption within DECnet-VAX 9 14 Support for simple project accounting 9 64 Lexical for getting RIGHTSLIST info 9 10 Provide callable OPCOM interface 9 13 Standardize format for printable output 9 69 Enhance MONITOR DISK 9 20 DCL return status enhancements 8 59 Enhance file access from user images 8 71 Provide more LAT accounting info 8 52 TPU should display all characters 7 1 Improve heterogeneous user group support 7 19 Improve DCLTABLES manipulations 7 58 Enhance COPY to copy ACL's 7 4 Provide SCS communication services 6 63 DECnet proxy access for SET HOST command 6 5 Abbreviated commands in DTR 6 56 Enhance DECnet remote file access 6 66 STB support on more devices 6 30 DEFINE/KEY enhancements 6 35 Provide F$FILE_ATTRIBUTES service 6 16 Support no CR/LF in DCL WRITE 6 42 Implement SYSGEN DISCONNECT 5 15 First/last file-specs in BACKUP log 5 VAX-93 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY SCIENTIFIC/ENGINEERING USERS Total ballots in this category: 207 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 101 24 Capture interactive session better 85 67 Individual field reports from AUTHORIZE 82 11 Queued requests for ALLOCATE command 69 71 Provide more LAT accounting info 62 13 Standardize format for printable output 58 19 Improve DCLTABLES manipulations 57 54 Send more developers to Symposia 56 2 Implement cluster-accessible tape drives 55 14 Support for simple project accounting 54 15 First/last file-specs in BACKUP log 53 12 Tape automatic volume recognition 52 6 More detail in VMS accounting records 52 26 Various VMSmail enhancements 51 42 Implement SYSGEN DISCONNECT 43 25 Enhance sysgen parameter readability 43 59 Enhance file access from user images 42 8 Improve VMS Mount service messages 41 20 DCL return status enhancements 40 1 Improve heterogeneous user group support 39 55 Provide sales update for customers 38 16 Support no CR/LF in DCL WRITE 38 64 Lexical for getting RIGHTSLIST info 37 58 Enhance COPY to copy ACL's 37 27 Enhance SET/HOST/DTE for more modems 37 35 Provide F$FILE_ATTRIBUTES service 36 5 Abbreviated commands in DTR 35 52 TPU should display all characters 35 18 Enhance command line editing 35 34 Combine DEFINE, ASSIGN commands 34 23 DCL /LOG needs more consistency 34 62 Security alarm messages to a file. 34 47 Resolve Fortran/Debug "TRUE" 33 37 $TRNLNM wildcard support 31 30 DEFINE/KEY enhancements 31 VAX-94 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY OFFICE AUTOMATION USERS Total ballots in this category: 151 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 81 24 Capture interactive session better 59 67 Individual field reports from AUTHORIZE 55 71 Provide more LAT accounting info 52 14 Support for simple project accounting 44 19 Improve DCLTABLES manipulations 41 54 Send more developers to Symposia 41 11 Queued requests for ALLOCATE command 39 13 Standardize format for printable output 39 6 More detail in VMS accounting records 39 16 Support no CR/LF in DCL WRITE 38 15 First/last file-specs in BACKUP log 37 12 Tape automatic volume recognition 37 59 Enhance file access from user images 36 1 Improve heterogeneous user group support 34 2 Implement cluster-accessible tape drives 34 64 Lexical for getting RIGHTSLIST info 33 55 Provide sales update for customers 33 5 Abbreviated commands in DTR 33 26 Various VMSmail enhancements 31 25 Enhance sysgen parameter readability 31 35 Provide F$FILE_ATTRIBUTES service 31 27 Enhance SET/HOST/DTE for more modems 30 30 DEFINE/KEY enhancements 30 62 Security alarm messages to a file. 30 58 Enhance COPY to copy ACL's 29 20 DCL return status enhancements 28 8 Improve VMS Mount service messages 27 37 $TRNLNM wildcard support 27 63 DECnet proxy access for SET HOST command 26 42 Implement SYSGEN DISCONNECT 26 18 Enhance command line editing 23 22 /USER for running detached jobs 22 61 End-to-end encryption within DECnet-VAX 21 41 Support for higher baud rates 21 VAX-95 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY TELECOMMUNICATIONS USERS Total ballots in this category: 89 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 28 Display open files on SHOW PROCESS 45 24 Capture interactive session better 38 67 Individual field reports from AUTHORIZE 34 11 Queued requests for ALLOCATE command 33 6 More detail in VMS accounting records 27 12 Tape automatic volume recognition 27 54 Send more developers to Symposia 27 13 Standardize format for printable output 27 26 Various VMSmail enhancements 26 8 Improve VMS Mount service messages 25 20 DCL return status enhancements 23 16 Support no CR/LF in DCL WRITE 23 55 Provide sales update for customers 22 5 Abbreviated commands in DTR 22 71 Provide more LAT accounting info 22 2 Implement cluster-accessible tape drives 21 64 Lexical for getting RIGHTSLIST info 20 15 First/last file-specs in BACKUP log 20 19 Improve DCLTABLES manipulations 20 56 Enhance DECnet remote file access 18 62 Security alarm messages to a file. 16 30 DEFINE/KEY enhancements 16 42 Implement SYSGEN DISCONNECT 15 1 Improve heterogeneous user group support 15 4 Provide SCS communication services 15 18 Enhance command line editing 15 61 End-to-end encryption within DECnet-VAX 14 23 DCL /LOG needs more consistency 14 35 Provide F$FILE_ATTRIBUTES service 14 25 Enhance sysgen parameter readability 14 59 Enhance file access from user images 14 27 Enhance SET/HOST/DTE for more modems 13 63 DECnet proxy access for SET HOST command 12 14 Support for simple project accounting 12 37 $TRNLNM wildcard support 12 VAX-96 PAGESWAPPER - January 1988 - Volume 9 Number 6 Fall 1987 US VAX Systems SIG SIR Ballot Results THE TOP 35 SIR's AS RANKED BY OTHER Total ballots in this category: 23 Fall 1987 Ballot SIR SIR Total Nr. Description Votes 71 Provide more LAT accounting info 14 28 Display open files on SHOW PROCESS 13 67 Individual field reports from AUTHORIZE 11 14 Support for simple project accounting 11 19 Improve DCLTABLES manipulations 10 54 Send more developers to Symposia 10 41 Support for higher baud rates 9 24 Capture interactive session better 9 66 STB support on more devices 9 27 Enhance SET/HOST/DTE for more modems 9 8 Improve VMS Mount service messages 9 1 Improve heterogeneous user group support 8 13 Standardize format for printable output 8 58 Enhance COPY to copy ACL's 8 6 More detail in VMS accounting records 8 15 First/last file-specs in BACKUP log 8 30 DEFINE/KEY enhancements 8 59 Enhance file access from user images 7 46 Standardize data-type support 7 3 Remove TU-58 boot requirement for 750 7 42 Implement SYSGEN DISCONNECT 7 12 Tape automatic volume recognition 6 20 DCL return status enhancements 5 33 DCL DEFINE enhancements 5 34 Combine DEFINE, ASSIGN commands 5 26 Various VMSmail enhancements 4 31 Add /BELL to various DCL commands 4 55 Provide sales update for customers 4 29 Enhance MOUNT/FOR for uninitialized tapes 4 52 TPU should display all characters 3 25 Enhance sysgen parameter readability 3 2 Implement cluster-accessible tape drives 3 56 Enhance DECnet remote file access 3 11 Queued requests for ALLOCATE command 3 22 /USER for running detached jobs 3 VAX-97 PAGESWAPPER - January 1988 - Volume 9 Number 6 INPUT/OUTPUT INPUT/OUTPUT A SIG Information Interchange A form for INPUT/OUTPUT submissions is available at the back of the issue. To register for on-line submission to the Pageswapper dial: (617) 262-6830 (in the United States) using a 1200 baud modem and log in with the username PAGESWAPPER. ================================================================ Note 585.25 Anyone use defrag programs? 25 of 26 "James H. Thompson" 46 lines 7-NOV-1987 23:03 -< Defrager tests >- ---------------------------------------------------------------- An electronic mail message from our systems programmer -- we were evaluating these programs only from the standpoint of how well they fit our needs. Subj: Thumbnail review of VMS disk defragmenters Diskeeper took one of our typical disks, used up all the contiguous free space trying to defragment some of the files, and decided that it was done. The fact that it didn't destroy the data on the disk would have only been a technicality if it had been a production run. The installation instructions do have a charming note that we should have used VMS BACKUP to reorganize the disk before running Diskeeper, but.... What we need is a tool that will fix our free space fragmentation problem. Having a program that might fix the problem but will probably turn it into a disaster and then quit is the pits. Defrag is just barely fast enough to be useful. It can do a so-so job of defragmenting about half the free space on an Eagle in one to two nights (of about four hours.) It can't be used in a daytime emergency because it uses a huge amount of memory (5 megabytes is much too small) and CPU time (about 45 minutes of CPU time before it even gets started.) It is better than VAX-98 PAGESWAPPER - January 1988 - Volume 9 Number 6 nothing, though. Rabbit looks fairly good. It defragmented both the files and free space reasonably well on an Eagle in a little over three hours (Defrag can't defragment the files on a live Eagle... it isn't fast enough to keep up with normal file activity by a factor of maybe ten.) I read the instructions for Diskit. There isn't any possible point in trying it. It runs offline, and you have to restore the disk if the system crashes or if you get a major disk error. To summarize... Diskeeper - NO. Defrag - probably not Rabbit - maybe Diskit- NO. James H. Thompson VeriFone Inc. 1100 Ward Ave, #760 Honolulu, HI 96814 808-536-1041 ================================================================ Note 585.26 Anyone use defrag programs? 26 of 26 "Alan E. Frisbie" 17 lines 9-NOV-1987 00:04 -< DiskKeeper NOT recommended >- ---------------------------------------------------------------- At the recent European DECUS symposium in Rome, there was a session that was supposed to evaluate the various disk defragmenters. What it ACTUALLY turned out to be was an hour-long commercial for DiskKeeper, given by their European representative. It was the most blatent display of commercialism that I have seen in 10 years of symposia. Unfortunately, the session chair had never chaired a session before and had not been properly briefed on what to do in such a situation. By the time that members of the symposium committee had been summoned and had declared that it was a commercial message, the session was almost over. VAX-99 PAGESWAPPER - January 1988 - Volume 9 Number 6 For this, if no other reason, I would NEVER buy DiskKeeper or any other product from this company. Alan E. Frisbie Flying Disk Systems, Inc. 4759 Round Top Drive Los Angeles, CA 90065 ================================================================ Note 660.33 Applications software standards 33 of 33 "Bob Huckins" 48 lines 6-NOV-1987 19:24 -< Comments from a Developer >- ---------------------------------------------------------------- I hope people are still watching this conference: I have some belated comments. As an OEM, I'm a little distressed at the paranoid tone of some of the suggestions. I understand (from experience) the guarded attitude that many people have towards layered product software, but some of the suggestions would make life difficult for OEMs and less sophisticated users. For example, the suggestion that privileged software be supplied in source form would increase the size of distribution kits (remember all those systems with RX50s!), would require one or more language processors on the target system, and would be ultimately futile, since who is going to look through tens of thousands of lines of code for trojan horses? Many of the other suggestions assumed that there is a guru managing the system, which is simply not true for the majority of VAXes running today. Another consideration for the development of standards is compatibility with older products. We have been shipping VMS-based layered products since VMS version 2, and our customers have developed procedures and programs which depend on our software being installed where it is (we patterned our software after DEC layered products, By the way: executables in SYS$SYSTEM, etc.). We (and our customers, I hope) would be willing to adjust to a new installation scheme (by this I mean not VMSINSTAL itself but where layered products are installed, how commands are defined, etc.), if DEC would put its stamp of approval on it, and it made installing all layered products easier. VAX-100 PAGESWAPPER - January 1988 - Volume 9 Number 6 If such a scheme were developed, I would like it to have these characteristics: 1. No editing of system startup file. Have some generic mechanism by which layered products can have an initialization procedure run automatically (at system startup time) by making an installation time request. 2. No manual tuning required. This can almost be done now by changing MODPARAMS.DAT, there needs to be a foolproof mechanism. 3. No editing of login procedures. As in note 1, a generic mechanism to invoke per process login procedures. As you can see, my concerns are more with unsophisticated users with small systems than with giant clusters managed by geniuses. P.S. Re Frank Nagy's note about multiple DCL command tables: I believe this was one of the top 10 SIR's several years ago but was quashed by DEC for performance reasons. Bob Huckins Nuclear Data Instrumentation Div. Golf & Meacham Rds. Schaumburg, IL 60196 312-884-3659 ================================================================ Note 663.32 Comments on the SPR process 32 of 43 "Kevin Angley" 18 lines 2-NOV-1987 17:54 -< DEC does care - but will they take action???? >- ---------------------------------------------------------------- In response to my SPR on RUNOFF which I returned for a more intelligent response, I got a call from a Peggy in Colorado who agreed that it was a stupid answer, and at least got me to believe that somebody at DEC cares that the SPR process sucks. So ... DEC does recognize this as one of their things to work on, and it would help them to help us if you made your complaints known through every avenue .. DECUS, these notes (they read them, you know), your salesperson, and probably the best bet is your Field Service manager. VAX-101 PAGESWAPPER - January 1988 - Volume 9 Number 6 Speaking from a company that does some degree of software support, I can tell you that one problem is that the brightest software types do not particularly like support work. There are a few rare exceptions that include ours and the reader (if the reader is in support work), but they are far and few between. I sympathize with DEC on this, but I don't think that is the whole problem. Kevin Angley 3301 Terminal Drive Raleigh, NC 27604 (919) 890-1416 ================================================================ Note 663.33 Comments on the SPR process 33 of 43 "Larry Kilgallen" 43 lines 2-NOV-1987 19:42 -< Physical Proximity of Supporters and Developers >- ---------------------------------------------------------------- Two hours ago I just went through a support contact in the other direction. Customer support got a question for which they did not have an answer and they asked me, as one of those who has written software the customer runs, how a particular aspect of the product worked with another vendor's software. On a much smaller scale, a situation much like DEC must handle for SPRs. The difference is that the vendor in question is *much* smaller, so that customer support and programming staff have face-to-face encounters every day (every 10 minutes, in fact). In a larger organization like DEC, this would not be possible, but if TSC were at the Spit Brook Road facility in New Hampshire at least the two groups would eat in the same cafeteria and see each other in the halls. Now I realize that DEC is trying to have a distributed *everything* (talking organizations, now, not computers), but I really think there is a time when international VAXnotes conferences are not adequate to make people all feel part of the same team. There is also the matter of developers in Reading England and other non-Spitbrook sites. I would propose that DEC put the technology in the phone system and let a switchboard operator in Colorado connect you with "The RPG Team" by switching you to support folks who are *colocated with the developers*. Then the RPG supporters can play volleyball against the RPG developers once a week and DEC can VAX-102 PAGESWAPPER - January 1988 - Volume 9 Number 6 still have it's distributed development environment. That seems to be the way it works for SNA Gateway products, but with a separate phone number to the building which houses support folks, developers, and DEC's collection of IBM machines. As far as cohesiveness in the support organization, at least some of DEC's technical writing is organized this way, with technical writers trained by a central group in the vagaries of SDML and then assigned to various development groups to do their actual writing. I am not sure who does performance reviews or anything like that, but manuals created in various locations still have a cohesive style. Larry Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 ================================================================ Note 663.34 Comments on the SPR process 34 of 43 "Dale E. Coy (505) 667-3270" 25 lines 2-NOV-1987 22:01 -< Integrated Support? >- ---------------------------------------------------------------- Amen to your comments, Larry. Particularly, the "one phone number" idea (which DEC has assured me for the last 2 1/2 years that they are "working on"). Don't know how many of you use the dialup DSIN, but my idea of fitting punishment for any DEC manager is to force him/her to use DSIN for a while. A sampling: 1. Listings in the "Flash Messages" sub-menu which, when selected, first say "getting messages" and then say "There are no Flash Messages for this product". 2. If you want to use Menu mode, you are first asked if you are using a VT100/200 terminal. If you say "Y", then DSIN does an inquire and, if you lied, they say so. With a VT340, that's annoying - but they should either inquire or believe me, not ask and then check. 3. This is now a Field Service system. Has been for months. I demonstrated it for my Field Service folks last week, and they were amazed that customers could access such a hi-tech thing. VAX-103 PAGESWAPPER - January 1988 - Volume 9 Number 6 DSIN is a fantastic idea. If they could just find somebody to write good software to support it..... DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 663.35 Comments on the SPR process 35 of 43 "Bill Mayhew" 43 lines 3-NOV-1987 14:54 -< If you or any of your IM force should be caught or killed... >- ---------------------------------------------------------------- re: co-location: Agreed, this is an excellent idea. At least _some_ products are currently "supported" this way. After escalating a series of unresponsive SPRs to the Digital product manager, I gained a fair amount of insight into how this product (at least) is supported. A couple of the highlights: 1. Unless the product is a high-volume product (e.g. VMS itself) it is likely that the CSC "support team" is actually the support team for a variety of other products -- potentially dozens of them. Needless to say, they don't know any one product well. This is, IN MY HUMBLE OPINION, counterproductive, since it's the more "obscure" products that one is MOST likely to need decent support for! 2. The machine resources at hand at the CSC are often inadequate. In many instances, the software product in question is not even installed on the at-hand system, and the service contact has to install the product, _then_ try to reproduce your problem. 3. While there is (at least for this product, which shall remain nameless to protect the ... well, let's say they're still on trial) in fact a group of support developers at the main engineering center, coresident with the product developers, the communication between the support portion of the development team and the CSC VAX-104 PAGESWAPPER - January 1988 - Volume 9 Number 6 support team is often fairly minimal. Unless one squeaks one's wheel fairly loudly, it is limited to Electronic mail message. If you squeak loudly enough, they actually talk to each other on the phone (after figuring out the time zone differences and compensating for lunch hours, etc.). 4. There is supposed to be a mechanism so that if a call is not responded to -- i.e. a fix or work-around or "restriction" noted -- with some period of time (I was told 7 days), then the problem is escalated to a CSC manager who gets involved with the engineering support people. In practice, if this happens, it is invisible to the customer. ("Transparent support" is not my idea of a good design criterion :-) ) While I appreciated seeing all this on the table, I was, needless to say, not amused. Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 ================================================================ Note 663.36 Comments on the SPR process 36 of 43 "Dale E. Coy (505) 667-3270" 25 lines 3-NOV-1987 18:04 -< A couple of hints about telephone support >- ---------------------------------------------------------------- Of course, there are measurements and standards for support (wouldn't you do it this way?) One of them is obviously "time between customer call and closure". Another one is probably "number of calls handled per support person". This leads to some very interesting gyrations. I don't know of any good way to measure quality, but I sure wish there was one. One little tidbit not usually recognized is that you don't have to accept "closure" of a problem from telephone support. Not that this should be used lightly, but if you get an answer that doesn't seem reasonable to you ("we don't understand what the problem is, can't duplicate it, and it will be fixed in a future release"), you can essentially demand that the problem be escalated to a manager, or even (horrors) a local office VAX-105 PAGESWAPPER - January 1988 - Volume 9 Number 6 referral. As an ABSOLUTE last resort, there is a DEC Customer Service Hotline. The folks there are described as sitting next to Ken at lunch. The number is jealously guarded. A very shrewd way to get it is to call your salesperson, say you're having no success getting your problem resolved, and ask for the hotline number. That'll get their attention, and may get you the number. If it doesn't, given a good enough war story, maybe the number will appear here. DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 663.37 Comments on the SPR process 37 of 43 "Bill Mayhew" 32 lines 3-NOV-1987 22:14 -< DECUS role >- ---------------------------------------------------------------- Larry made a comment about DECUS involvement in "fixing" the SPR process... Philosophically, I think the problem is not unlike that of curing an alcoholic or drug addict: he/she has to _want_ to be cured, first. (Any other interesting parts of that analogy are left as an exercise for the reader.) Some time ago, the various SIGs used to reprint SPRs filed by their members and Digital responses. This has the effect of making the process' shortcomings "see the light of day" rather than be forever condemned via the infamous "This SPR will not be published" smokescreen. I think there was even some discussion here about revitalizing that approach. Personally, I think it has some merit, but only if _enough_ of us copy our SPRs and responses to the relevant DECUS newsletters, and only if the relevant editors are prepared to publish 'em. The solution we'd all love to see, of course, is for Digital to stand up and say "Hey, we know there are problems with the SPR system (and with software support in general), here's what we're doing about 'em" -- or even, "Hey, we understand _you think_ there are problems, what would you _like_ us to do about 'em?". VAX-106 PAGESWAPPER - January 1988 - Volume 9 Number 6 However, I see no evidence of this. I see good intentions at the CSCs, backed up by insufficient resources, local office management that does not have the clout or the mechanisms to get problems resolved unless you're a Major Account, and general erosion around the edges of Digital's hard-won image as a top-flight system support provider. I'd love to see some DECUS leaders have a serious meeting with some Digital software support managers about this, but I'm not optimistic it would help. We've (all of us) been complaining about this too long; it seems clear that Digital does not take these complaints seriously. Maybe the "guerrilla warfare" tactic of published SPRs is our best next bet. Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 ================================================================ Note 663.38 Comments on the SPR process 38 of 43 "Jack Patteeuw" 5 lines 4-NOV-1987 10:30 -< Get problem number ! >- ---------------------------------------------------------------- re:.35 item 4) Automatic escalation of problem Don't **EVER** hang up the phone after talking to CSC without getting a problem number !! If it's not logged then it won't be traced/followed up ! Jack Patteeuw Ford Motor Co. Electrical and Electronics Division 31630 Wyoming Livonia, MI 48150 313-323-8643 VAX-107 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 663.39 Comments on the SPR process 39 of 43 "Bill Mayhew" 9 lines 5-NOV-1987 10:38 -< Whose job is this, anyway? >- ---------------------------------------------------------------- re: .38, getting the problem number Of course. I always do that. That doesn't mean, however, that it is followed up and escalated "automatically", the way it's "supposed to be". Sure, I can call them, number in hand, and say "WTH's my problem fix?"... but I have this deep-seated aversion to paying Digital for allowing me the privilege of doing their job for them. Must be my upbringing, or something in my water... Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 ================================================================ Note 663.40 Comments on the SPR process 40 of 43 "Bruce Bowler" 32 lines 5-NOV-1987 16:21 -< The loop >- ---------------------------------------------------------------- I had a long chat yesterday with an SPR manager (she shall remain nameless to protect her sanity) who finally described for me what happens to an SPR. First, you send it to Maynard, they transcribe it onto their system, which sends it to Colorado Springs, where they weed out duplicates, try to solve it, and then send it back to development. As an alternate, you can enter it on DSIN, which goes through essentially the same steps except in starts in Colorado rather than maynard. But, SPR's from DSIN are not considered time critical so are only uploaded to SPR team every 3-4 days. (she assured me that this was going to change to a more time critical event) VAX-108 PAGESWAPPER - January 1988 - Volume 9 Number 6 As a final alternative, you can call Colorado and request that the support person you are talking to submit an SPR for you. This is the fastest way to get an SPR into the system. One suggestion that I made to them was for them to periodically search their database for unanswered SPR's and send out a postcard telling us that they have not forgotten. I have an SPR that I sent in 2 years (24 months!!!) ago that I have not received an answer for. But yes they are still working on the problem. Finally, don't be too quick to blame the developer/maintainer for lousy SPR answers. They write the original answer which then gets passed through a long chain of hands (including lawyers) before you see the answer. They may have said "patch xyz.exe as follows" and what you get is "ûfixed in a future release of VMS after version 4.6". The developers/maintainers are really on our side, in many cases they want what we want, but the #%$&#^@%#$ lawyers get in the way. Bruce Bowler General Electric 1 River Road Bldg 2 Room 609 Schenectady, NY 12345 ================================================================ Note 763.15 VT320 Character Cell Terminal 15 of 16 "Bill Mayhew" 31 lines 15-NOV-1987 19:35 -< Further user evaluations >- ---------------------------------------------------------------- Having finally gotten the VT320-DAs I ordered on announcement day (from an alternate source after it became clear that my distributor had botched the order and was trying to cover it up by blaming it on DEC)... I have to agree that the character presentation is less than stellar. I haven't done a side-by-side comparison with the 220. I don't particularly find the 320 "objectionable", but it is definitely a disappointment. My impression is that the increase in resolution was not sufficient to compensate for the larger screen size, with the result being that the characters appear "dotty" (sort of like "grainy") in the vertical dimension. This may also be related to the size of each individual dot. VAX-109 PAGESWAPPER - January 1988 - Volume 9 Number 6 The monitor unit also has, overall, a substantially "cheaper" feel to it. This is probably no surprise, considering the price, but it is out of character (ahem) for DEC. I would have rather seen a more solid product at a price between the 320 and the 220, or a product that offered slightly reduced functionality at a real barrel-bottoming price tag. The terminal is not, again, "bad"... but the overall impression is of a Wyse or something, with DEC styling and the solid DEC keyboard. Also of some interest: I note that the copper shielding which, I thought, lined the 220 cabinet is absent in the 320, and that the 320 only has an FCC class A rating (I thought the 220 had a B, but again I may be mistaken). Also, about half the units I received have been reported to exhibit a very slight "bowing" of the horizontal scan line, so far noticeable only in Set-up, and possibly something that will go away as the units either burn in or just with normal daily warm-up. As I said, it's _very_ slight. I'm waiting to get my hands on one of these specific units before reaching any conclusions, and warranty calls to DEC may or may not become involved. Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 ================================================================ Note 763.16 VT320 Character Cell Terminal 16 of 16 "Kevin Angley" 6 lines 16-NOV-1987 15:51 -< This comedy doesn't cost you a dime >- ---------------------------------------------------------------- The previous note complained of two things on his amber-phosphor VT-320 - a grainy character presentation, and a status line that was undulating. Therefore, you have Amber waves of grain. Kevin Angley 3301 Terminal Drive Raleigh, NC 27604 (919) 890-1416 VAX-110 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 784.17 In search of a modem 17 of 24 "John K. Doyle, Jr." 21 lines 20-NOV-1987 20:20 -< A vote AGAINST DEC datacomm products >- ---------------------------------------------------------------- I have used most of DEC's datacomm products at one time or another, starting with the DF02. I have found them to be overpriced, incompatible with the rest of the "world" (and at times, each other), and unreliable. I have basically given up on DEC for anything that can be wired directly into the telephone network. I presently have 8 DF224 scholars (with the original ECO that gets rid of the line of UUUUUUUU's problem) and a pair of DFM-16 multiplexers with integral 9.6kbps modems. I have had UNTOLD numbers of hardware failures with the DFM's and we have completely replaced one of them at least twice in one year. What's really annoying about this is that Field Service generally does NOT stock a lot of "datacomm" spares and my regional distribution center (Santa Clara) doesn't either. My DF224's need to be powered off and back on from time to time as well in order to clear "hangs". This hang condition even occurs on my DF224 at home which is connected to a Rainbow and sits idle most of the time. Finally, a friend of mine who runs a "Call 911" emergency response dispatch system recently ordered and installed some of the new MUXserver-100 boxes. They MUXserver seems like a great idea. She went through three (3) of them before getting one that worked. Any similar experiences out there or am I just having a really long streak of bad luck? John K. Doyle, Jr. Steiner, Levi & Co. 2550 Mercantile Drive Suite C Rancho Cordova, CA 95670 (916) 638-2600 VAX-111 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 784.19 In search of a modem 19 of 24 "Dale E. Coy (505) 667-3270" 38 lines 21-NOV-1987 00:20 -< Discussion of VA2400s >- ---------------------------------------------------------------- | < Note 784.3 by NODE::US129532 "Larry Kilgallen" > | | -< A vote against Racal-Vadic (the company) >- | | The 2400 baud modem I need is one which handles the Vadic 3400 | modulation scheme as well as Bell 212A. The Vadic 3400 modulation scheme (2250/1150 Hz) was invented so that you could do 1200 baud from an acoustic coupled modem. That's the only advantage it has over the 212 scheme (1200/2400). I haven't seen an acoustic coupled modem in use for years (lots of them on shelves, though). My understanding is that Vadic has nothing at 2400 which will also handle the 3400 modulation at 1200 baud. I understand your frustration at not being able to get an answer, but: Unless you have some acoustic coupled VA345x modems in use, calling a VA345x from a VA2400 works fine. Likewise, calling a VA2400 from a VA345x works fine. The VA345x will automatically recognize that it needs to be in 212 mode. In other words, I still recommend the 2400. This connection to Pageswapper is made via 2400PA (non-MNP version) to the rack-mount equivalent of the 2400PA in our computer room (these support MNP), and then out via another 2400 card to your system. No problems whatsoever (if Ma Bell does a good job). Note that the connection from our VAX to Pageswapper TRIES to use 2400 baud and TRIES to use error correction - and then gives up on both gently. I just get a plain old Username: prompt. Not so on the OASIS system, which uses DEC modems (3 at 1200 and 1 at 2400). I sometimes have to try 2 or 3 times if I call at 1200. I have given up trying to call at 2400 baud because I always get one login failure, the Username: prompt, and am left with about 15 characters I have to ^U before I can enter my username. I'm sure it's the interaction with the error correction inquiry from my end, but it's so annoying that I just put up with 1200 baud. VAX-112 PAGESWAPPER - January 1988 - Volume 9 Number 6 DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 784.20 In search of a modem 20 of 24 "Larry Kilgallen" 30 lines 21-NOV-1987 08:20 -< Someone out there might require Vadic modulation >- ---------------------------------------------------------------- | Note 784.19 Unless you have some acoustic-coupled Vadic 345x... My problem is that the site in question is *this* site and I am trying to ensure that *any* 1200 baud modem in the hands of a DECUS member will work. I am certain I have seen in catalogs of yore some non-acoustic (electric?) modems from Vadic which have Vadic modulation as their only 1200 baud modulation scheme. I do not want to exclude somebody who happens to have one of those. (Thank you for the information about the relationship of the protocol design to acoustic coupling, though, I had never heard that before.) I share your degree of enthusiasm over DEC modems. My particular situation is trying to call the Electronic Store (no I have not forgotten about them, just about the chance of ordering anything from them). Using Vadic 3451-PA modems (from several sites) I find it takes up to 20 tries to get a connection, and that is without any problem of busy lines. I have no trouble with Vadic 3451-PA dialing various client sites and connecting with Bell 212A modulation, so lacking any details I chalk it up to the supposition that the Electronic Store must be using DEC modems of some flavor. I lack your technical prowess at telling what a modem is doing to the other modem, but deepest soul-searching has not allowed me to pin the modem handshake problem on the fact that the Electronic Store seems to run All-In-1! Larry Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 VAX-113 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 784.24 In search of a modem 24 of 24 "Chris Erskine" 25 lines 24-NOV-1987 07:31 -< MAJOR advantage to 3400 >- ---------------------------------------------------------------- | The Vadic 3400 modulation scheme (2250/1150 Hz) was invented so | that you could do 1200 baud from an acoustic coupled modem. | That's the only advantage it has over the 212 scheme | (1200/2400). I believe that the 3400 scheme was announced before Bell announced the 212A. There is a much greater advantage that the 3400 has over the 212A. Bell created a modem which in the third harmonic of the transmit frequency (I think and also depending on which of the modems you are looking at) falls in the receive range. This means that if you have any sort of impedance mismatch, start to see trash on your terminal. I have had a number of times using the same lines that a 212A would not work but the 3400 would. If I had my choice for a 1200 baud protocol, I would take the 3400 every time. Too bad Ma Bell 'does it better' and the 212A protocol became the real standard. I have fought many times with Bell trying to tell them that their voice grade lines did not meet tariff because I could not get a Bell 212A to work on it. (Voice: You need to talk to Data. Data: That is a voice grade modem, you need to talk to the Voice folks...) Chris Erskine 23 S Holcomb Clarkston, MI 48016 (313) 524-8836 VAX-114 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 818.0 MicroVMS V4.6 Bug for DHV11s 1 reply "VICTOR LINDSEY" 7 lines 9-NOV-1987 22:48 ---------------------------------------------------------------- DHV11s on MicroVAXs have a new annoying feature since the installation of MicroVMS V4.6 to wit: if you issue a Ctrl/C or Ctrl/Y while a listing is in progress, you run a high risk of "hanging" the process. DEC has traced the bug to SYS$SYSTEM:YFDRIVER.EXE. In the meantime, do not use the /DMA on your "$ SET TERM..." commands in your SYSTARTUP.COM file--instead, specify /NODMA. VICTOR LINDSEY VLSystems, Inc. 2691 Richter Ave., Suite 118 Irvine, CA 92714 714/660-8855 ================================================================ Note 818.1 MicroVMS V4.6 Bug for DHV11s 1 of 1 "Jack Patteeuw" 4 lines 11-NOV-1987 16:19 -< A better solution >- ---------------------------------------------------------------- see note 760.7 and 802.1 A better solution is to use YFDRIVER from 4.5. It's working fine on my system for two months now ! Jack Patteeuw Ford Motor Co. Electrical and Electronics Division 31630 Wyoming Livonia, MI 48150 313-323-8643 VAX-115 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 820.0 Foreign Disk Comments 6 replies "Ken Robinson" 10 lines 12-NOV-1987 08:09 ---------------------------------------------------------------- Our sight is getting some non-DEC (SI and/or Emulex) RA-81 compatible disk drives. Anyone have any comments or warnings about them. Ken Robinson Bell Communications Research 444 Hoes Lane, Room 4d449 Piscataway,N.J. 08854 (201)699-8796 ================================================================ Note 820.1 Foreign Disk Comments 1 of 6 "Brian Tillman, Smiths Industries." 2 lines 12-NOV-1987 08:55 -< Psst... >- ---------------------------------------------------------------- I've heard rumors that SI is bad and Emulex is good. No first hand knowledge, though. Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 ================================================================ Note 820.2 Foreign Disk Comments 2 of 6 "Bob Hassinger" 10 lines 12-NOV-1987 15:34 -< Not happy with Emulex - reputation going down hill.. >- ---------------------------------------------------------------- There has been discussion elsewhere recently (DECUServe?) to the effect Emulex used to be a good company with good products but they have gone down hill more recently. VAX-116 PAGESWAPPER - January 1988 - Volume 9 Number 6 I know I have had very unsatisfactory experience with them on the SC7000 and do not care to consider their products in the future as a result because I am not convinced I can depend on them to be the kind of company I need. Bob H Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 820.3 Foreign Disk Comments 3 of 6 "Terry Kennedy" 18 lines 13-NOV-1987 03:05 -< Never another Emulex! >- ---------------------------------------------------------------- I wouldn't ever buy anything from Emulex again - I bought an SC31 (Unibus RP/RM disk emulation) and a TC02 (Q-bus MS tape emulation). I was promised a free upgrade to MSCP support for the SC31 when it became available. Later, I saw an ad for the SC41 (a MSCP'd SC31). My sales representative's response was 'Oh, we decided to make it a different model product - sorry'. We never could get the TC02 to properly emulate a TSV05 tape - it always thought it was a broken TK25 - so they finally gave us our money back on it. They may do better in other product areas but I don't want to find out. A good guideline to follow is to ask 'Does this controller require patches or a special driver to operate?' If the answer is yes, run far and fast - it *will* bite you in the future! If the answer is no, get it in writing before signing the check. I suggest this with any vendor. Terry Kennedy 95 Mohawk Trail Ringwood, N.J. 07456 (201) 435-1890 VAX-117 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 820.6 Foreign Disk Comments 6 of 6 "JEFF KILLEEN" 23 lines 14-NOV-1987 10:56 -< BUYING EMULEX? - I ALSO HAVE SOME BRIDGE PROPERTY..... >- ---------------------------------------------------------------- Over the last 3.5 years I have had the following problems with Emulex TC02 - won't boot, won't fully support TC05 emulation - charged me to fix that one at $225 per controller. SC03 - Did not work right with 11/73 they fixed that one. UC02 - Took them 5 months to get it working - meanwhile I had to hold a $50,000 system in inventory TC05 - Has never worked reliably The biggest problem with Emulex is their FCO policy. You have to pay for all FCO's after the first year even if their board firmware was in error. Before you buy Emulex hardware ask them to answer this question in writing. If I have had your board for over a year and a new version of the operating system has a problem with your board because of a bug in the firmware or because they did not properly emulate the DEC device who pays for the new firmware? JEFF KILLEEN 31 HOPEDALE ST. HOPEDALE, MA. 01747 617-478-8098 VAX-118 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 821.0 Undocumented V4.6 Print system fix. 9 replies "JIM PALMER" 42 lines 12-NOV-1987 12:27 ---------------------------------------------------------------- There appears to be an undocumented fix in the V4.6 print queue facility. In summary, going from v4.5 to V4.6 changes the sense of the /TRUNCATE qualifier used with the DEFINE/FORM DCL command verb. Under V4.5 the system defined DEFAULT form (#0) is explicitly defined with the /TRUNCATE qualifier. Under V4.6 the definition remains unchanged. However, under V4.5 the action of this switch appears to have been inverted. That is, the print symbiont appears to interpret it in the reversed sense, (ie. /NOTRUNCATE). Under V4.6 the action is correct. I encountered this problem after upgrading a stable system to V4.6 and found that users printing files with a word processing package experienced what first appeared to be random truncation of records within their files. [1] On closer inspection with DUMP, the records affected were all greater than the /WIDTH value, because they were filled with escape sequences to control our LN03. I have found no reference to this apparent fix in the release notes. Would be interested if others have experienced the same. In order to prove a possible source of the problem, I reached for my V4.6 fiche set. However, I found that it hasn't shipped yet!!!!. Has anyone else (who bought the product) received their V4.6 fiche kits? [1] On our V4.5 system the 'wrong' qualifier action was 'right' for our application. Thus, the old system appeared correct. JIM PALMER 3 BROOKDALE IRVINE, CA. 92714-3338 (714) 458-3028 VAX-119 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 821.1 Undocumented V4.6 Print system fix. 1 of 9 "Bob Hassinger" 12 lines 12-NOV-1987 15:42 -< What the (bleep) IS going on with the fiche? >- ---------------------------------------------------------------- | .. Has anyone else (who bought the product) received their V4.6 | fiche kits? This is getting to be a significant issue and I wonder what is going on! DEC told us we had to order the fiche separately, we went through a great deal of difficulty placing the order and getting it processed, we gave them the money and now we do not get what we paid for! What the (bleep) is going on? Bob H Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 821.2 Undocumented V4.6 Print system fix. 2 of 9 "JEFF KILLEEN" 4 lines 12-NOV-1987 17:53 -< ITS COMING >- ---------------------------------------------------------------- | What the (bleep) is going on? The *PROPOSED* FCS date is 16-NOV-87 JEFF KILLEEN 31 HOPEDALE ST. HOPEDALE, MA. 01747 617-478-8098 VAX-120 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 821.3 Undocumented V4.6 Print system fix. 3 of 9 "Terry Kennedy" 13 lines 13-NOV-1987 03:10 -< Beep Beep! >- ---------------------------------------------------------------- They must be hand-writing each slide under a microscope - after all, the code *has* been shipped - what's to change? As a matter of fact, since the mandatory update patches never make it into the fiche, the fiche could be ready well *before* the VMS kits are shipped. Doesn't DEC realize that many of us depend on the fiche for patching device drivers and can't install the release without it? Perhaps now is the time for all of us to make a concerted 'comment' to our sales/support reps about this. The 'future unannounced major release of VMS' will probably break all of our drivers - do we really want to wait N months for the fiche for it? Terry Kennedy 95 Mohawk Trail Ringwood, N.J. 07456 (201) 435-1890 ================================================================ Note 821.4 Undocumented V4.6 Print system fix. 4 of 9 "Bob Hassinger" 17 lines 16-NOV-1987 11:35 -< Perhaps DEC knows EXACTLY what the result of the delays is... >- ---------------------------------------------------------------- | .. Doesn't DEC realize that many of us depend on the fiche for | patching device drivers and can't install the release without | it? Just an observation - DEC may be aware that certain third party vendors need to patch the drivers they provide to their customers (say, for example Emulex and the drivers for their disk systems). If access to new releases and fiche are delayed it puts such companies at a competitive disadvantage because their customers can not update to the new release until they have the patched drivers for their third party hardware. VAX-121 PAGESWAPPER - January 1988 - Volume 9 Number 6 (I was VERY careful to be sure my third party disk system would run with UNMODIFIED drivers to avoid getting caught in this trap). Bob H Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 821.5 Undocumented V4.6 Print system fix. 5 of 9 "Alan E. Frisbie" 12 lines 16-NOV-1987 22:05 -< Just because you're paranoid... >- ---------------------------------------------------------------- | | Just an observation - DEC may be aware that certain third party | vendors need to patch the drivers they provide to their | customers (say, for example Emulex This may be excessively paranoid, but have you ever considered the possibility that DEC might carefully examine 3rd party controllers for imperfect emulations. They could then modify the drivers so that they would work with DEC devices, but not with the 3rd party ones. I seem to recall that they did this with the Intersil PDP-8 look-alike chip. Alan Alan E. Frisbie Flying Disk Systems, Inc. 4759 Round Top Drive Los Angeles, CA 90065 VAX-122 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 821.6 Undocumented V4.6 Print system fix. 6 of 9 "Larry Kilgallen" 15 lines 16-NOV-1987 22:19 -< Who at DEC has the time? >- ---------------------------------------------------------------- Individuals at DEC have their hands full examining DEC devices and trying to make drivers that *will* work with *them*. I cannot imagine a technical person at DEC going to that trouble of their own accord (even with employee stock purchase plans, their interests are not *that* tied to the company interests). For a manager to order a technical person to do that would be highly dangerous, leading to restraint of trade suits, etc. if the word ever leaked. I, for one, am confident that technical work against third parties is confined to keeping new DEC methods secret, not cutting them out after the fact. Larry Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 ================================================================ Note 821.7 Undocumented V4.6 Print system fix. 7 of 9 "Dale E. Coy (505) 667-3270" 21 lines 17-NOV-1987 02:08 -< "Good-old-days" Warstory >- ---------------------------------------------------------------- Back in the "good old days", when logic was really expensive, controllers were fairly simple and well documented. DEC would document some bits of the registers as write-only, some as read-write, and some as read-only. Some of the read-write bits were only written to by the drivers. 3rd party vendors, to save money, would implement them as write-only. I remember 3 separate instances where a new version of RSX started writing and reading a bit, and some luckless 3rd party vendor had to modify all hardware around the world. I also remember 1 instance where a bit was "suddenly" discovered to be read-write, in spite of the documentation. Now, I wouldn't accuse someone of doing this deliberately, but it sure was a coincidence that it happened so often. VAX-123 PAGESWAPPER - January 1988 - Volume 9 Number 6 Of course, we're all much more sophisticated today. Murphy's Law being fully operative, some 3rd party things will break if DEC just proceeds in a routine fashion. Heck, some DEC things break when a new version of VMS comes out. DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 821.8 Undocumented V4.6 Print system fix. 8 of 9 "Bob Hassinger" 35 lines 17-NOV-1987 12:02 -< DEC has done it in the past for sure - Warstory >- ---------------------------------------------------------------- I agree that in general today DEC should not and probably does not *intentionally* put traps in their hardware and software to make life harder for 3rd party vendors. I also agree that over the years the normal design process has provided plenty of problems for third parties without any special effort on DEC's part. That does not preclude the possibility that someone at DEC did not feel compelled to go out of his way to get the microfiche out on time even given the impact a delay would have on third parties, some of whom DEC is not getting along with too well just now. That is not a technical issue and can happen on the administrative level. That said, let me tell you from personal knowledge that no matter what the practice is today, DEC most definitely did take steps to attack a third party vendor in at least one case in the past. This is FACT: In one release of OS/8 there was code that explicitly looked for a 3rd party clone of the PDP-8 that was being marketed at the time. If this version of OS/8 found it was running on one of these machines it not only refused to work but it also went out and intentionally erased the system image on the disk! VAX-124 PAGESWAPPER - January 1988 - Volume 9 Number 6 This was known to a number of people. I recall a public discussion in a DECUS session where John Alderman and I beat on the DEC people present, making the point that not only was this a BAD policy but John estimated it would take a good high school kid a couple of days at most to track down the trap and patch it anyway. In fact, in the next release the trap was gone and the *very* scarce space it had used was devoted to other, more productive capabilities. Bob H Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 821.9 Undocumented V4.6 Print system fix. 9 of 9 "John Osudar" 7 lines 17-NOV-1987 20:26 -< DEC doesn't need to trip up 3rd parties >- ---------------------------------------------------------------- From what we've seen of at least a couple third party disk vendors, they don't need DEC's help to make their hardware stop working the way the salesperson claimed it was supposed to... but I do wonder why it takes over two months to get the fiche out. (Yes, I know that the scheduled FCS date was Monday; I'll believe it when that little (?) brown envelope is sitting on my desk, next to my VAXstation and portable fiche reader.) John Osudar Argonne National Laboratory 9700 S. Cass Ave. Bldg. 205 A-051 Argonne, IL 60439-4837 (312) 972-7505 VAX-125 PAGESWAPPER - January 1988 - Volume 9 Number 6 ================================================================ Note 824.0 Questions on Oracle 6 replies "Jack Patteeuw" 15 lines 12-NOV-1987 17:11 ---------------------------------------------------------------- Another site in our company has a VAX running Ultrix (why I don't know!) and is connected up to our network with DECNET/Ultrix. This system also has Oracle. It is being proposed that Oracle be purchased for one of our VMS machines so that data can "shared" across the network (these machines are in different DECNET areas if that matters). Does anyone have any first hand knowledge of whether or not Oracle can "share" data across DECNET with either similar or dissimilar operating systems ? I have heard a few horror stories about site that have Oracle, would anyone care to share their experiences (good and bad) from both system managers as well as users perspective. Jack Patteeuw Ford Motor Co. Electrical and Electronics Division 31630 Wyoming Livonia, MI 48150 313-323-8643 ================================================================ Note 824.1 Questions on Oracle 1 of 6 "M. Erik Husby" 25 lines 13-NOV-1987 09:32 -< Comments on Oracle >- ---------------------------------------------------------------- We have used Oracle across VMS DECnet's with no problems. If Ultrix/DECnet is compatible with VMS DECnet there should not be any problems. It can be set up so that it is transparent to the Oracle user where the data is. Either through logical names or DCL symbols can be used to direct the Oracle login to the remote system. VAX-126 PAGESWAPPER - January 1988 - Volume 9 Number 6 The remote system will be running a DECnet server task that handles the queries and ships back the data. Under VMS one can define this server in many different ways so that your security needs can be met. As for horror stories -- Oracle does work the way it is advertised. We have been using them for years and are reasonably satisfied. Be sure that you have plenty of cpu horsepower on the system where Oracle will be doing its work. Plenty of memory is a good idea as well. One comment -- If you plan on writing application programs that call Oracle directly be prepared to rebuild Oracle and those application programs on every VMS upgrade and update. Oracle is built using position dependent code and privileged shared images. If the VMS runtime libraries change, which they usually do with a new release, then any position dependent programs need relinking. M. Erik Husby Project Software & Development 14 Story St. Cambridge, MA. 02138 (617)-661-1666 ================================================================ Note 824.4 Questions on Oracle 4 of 6 "M. Erik Husby" 10 lines 17-NOV-1987 14:38 -< Remote queries and updates of 1 database >- ---------------------------------------------------------------- | | | The way you described it, Oracle does not really "share" data (a | la DEC's distributed database manger/Rdb) but does allows remote | queries (sort of like DTR does), correct ? You can have remote queries and updates of one remote data base. You currently can not join two data bases -- this is version 5 oracle. I will have to check with my Oracle Guru to determine if that feature is coming in Version 6. M. Erik Husby Project Software & Development 14 Story St. Cambridge, MA. 02138 VAX-127 PAGESWAPPER - January 1988 - Volume 9 Number 6 (617)-661-1666 ================================================================ Note 824.5 Questions on Oracle 5 of 6 "Alan B. Hunt" 108 lines 23-NOV-1987 18:42 -< some info >- ---------------------------------------------------------------- Some or all of this may be old news to Jack since we work together but I will pass along some of the things we found in researching Oracle and Rdb. We wanted Rdb (users too) but the corporate standard is Oracle and unless there was a true technical reason they wouldn't let us deviate. Here is some of what we found: 1. We finally went Rdb because it was found that V5.x of Oracle (still the current release) would not work with RS/1 put out by BBN if you wished to retain the ability to write programs which called RS/1 through its shareable image interface and also wanted to use the RS/1 - Oracle interface from BBN. You had to install RS/1 as a shared program not a shareable image. The reason is that ORACLE uses a BASED shareable image which has address conflicts with RS/1. ORACLE supposedly has come up with a work around for the lifetime of V5 using mailboxes. It is not a normal supported option, however. Their real fix is Oracle V6 which eliminates the BASED shareable image and goes to a standard shareable image. This will also hopefully eliminate the need to relink ORACLE and user programs linked to it after each VMS update. 2. If you read the manuals, you will find if using the language preprocessors and a high level language to interface to the database you must turn off the optimizer because ORACLE may have problem finding variables if the optimizer moves them to a register. We have found the optimizers to be quite effective in reducing memory and cpu usage and did not care for this. 3. Oracle is written in C. To use the language calls to Oracle they require you to conform to the C calling standard for passing parameters. Therefore you pass by value and do not use descriptors for strings. This seems awkward for the VMS environment unless you VAX-128 PAGESWAPPER - January 1988 - Volume 9 Number 6 program in C. 4. Oracle tends to require you to create one physical database and all logical databases go into it. Therefore a database manager is required to coordinate it. (This was confirmed by other Oracle users in the company.) Rdb on the other hand allows any user to create his own physical database and not worry about others or a database manager. (Some might not like this but we did.) Oracle also does not support volume sets although it seems to indicate it does. More than one site we talked to said they had tried it but that it definitely would not work. Oracle does not do dynamic database expansion. The database manager must expand it manually. Just make it big when you start is the work around. You can have an Oracle database span two or more disks so large databases are possible. 5. Each physical database requires four detached processes to be active to support it. If you have two physical databases, then eight processes are required and so on. In a cluster these must exist on every machine accessing the physical database. 6. For VMS V4.0 they were 6-9 months late coming out with a database that would run. Cluster support (where two machines could access the same database) came even later. 7. Cost was high. In a cluster the second machine was more than the first from what we could tell. (They do not like and discouraged us from running it in cluster mode.) 8. Oracle does not use VMS identifiers or usernames. Each user (unless a wide open database is wanted) must be given an Oracle username and password. The Oracle username can be associated with a VMS username so Oracle can allow a user into Oracle without entering the Oracle username. You can't easily use UIC groupings or general identifier groupings in Oracle to grant or deny access. VAX-129 PAGESWAPPER - January 1988 - Volume 9 Number 6 9. Since their aim is to keep Oracle portable they decided not to use the VMS Distributed Lock Manager for locking. Rdb does and I believe Ingress does as well. On a single machine this was not a big factor since one of the detached processes described above maintains the locks in a global section in memory. This is similar to what VMS does for its lock manager. In a cluster, however, this posed a problem since memory is not shared. They use a area on disk to keep a copy of what is is in memory and after every lock change or when a update is needed the lock table is copied from memory to disk or disk to memory. According to what they said you would suffer a slight penalty. The information they gave us was as follows: VAX CLUSTERS Performance Impact Extra read for each table in a query Two or three extra writes for each table in a transaction Extra read for each statement parsed Impact on elapsed time is small for operations on many rows Impact on elapsed time is moderate for operations involving only a few rows, or where commits occur frequently Impact on CPU time is small We know of no plans for Oracle to use the VMS Distributed Lock Manager. On the other hand I would say that today Oracle on a single machine is probably faster then Rdb but slower than Ingress from what we have heard. We did no benchmarks since speed was not the primary issue. The user interface is quite powerful and well developed from what others in the company say. VAX-130 PAGESWAPPER - January 1988 - Volume 9 Number 6 It is portable between a large variety of machines. Oracle V6 is suppose to fix many of the maintenance headaches and will allow after image journaling in the VAX Cluster environment. Currently it is not supported in clusters. ALAN B. HUNT 26803 BERG RD. #301 SOUTHFIELD, MI 48034 ================================================================ Note 826.0 Error in newsletter article No replies "Alan E. Frisbie" 37 lines 13-NOV-1987 19:38 ---------------------------------------------------------------- The November, 1987 issue of the SIG Newsletters arrived today. In the article, "Migration to MicroVMS from DEC Operating Systems", there is an error that should be noted before it causes somebody some grief. On pages VAX-8/9, is the following exchange: Q: I've programmed on RSX for a while in MACRO and I have to convert to a MicroVAX... I haven't looked at all at MACRO in VMS. Is it much different? A: Totally different. You are going to get very confused, very fast. Moving R0 to R1, for example, works in reverse. It will take R1 and put it in R0, not R0 into R1. THIS IS TOTALLY WRONG!!! The direction of a MOVe instruction is EXACTLY THE SAME as on a PDP-11. Anyone who thinks it is reversed is going to get very confused, very fast. I have found it relatively easy to convert MACRO-11 to MACRO-32, as long as the original PDP-11 code was easy to understand. Even RSX I/O drivers can be successfully converted. For more information on converting MACRO-11 to MACRO-32, consult the series of articles published by the IAS SIG. They are written by Frank Borger of the Michael Reese Medical Center. The first installment was in the August, 1986 issue of "The DeVIAS Letter", the IAS SIG newsletter. VAX-131 PAGESWAPPER - January 1988 - Volume 9 Number 6 Alan E. Frisbie Flying Disk Systems, Inc. 4759 Round Top Drive Los Angeles, CA 90065 VAX-132 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX System SIG Committee List VAX System SIG Committee List As of June 24, 1987 CHAIR (CORE) Susan T. Rehse Lockheed Missiles & Space Co. O/19-50, B/101, P.O. Box 3504 Sunnyvale, CA 94088-3504 VICE-CHAIR (CORE) WORKING GROUP COORDINATOR Ross Miller Online Data Processing, Inc. N 637 Hamilton Spokane, WA 99202 SYMPOSIA COORDINATOR (CORE) Jack Cundiff Horry-Georgetown Technical College P.O. Box 1966 Conway, SC 29526 COMMUNICATION COORDINATOR (CORE) Don Golden Shell Oil Company Westhollow Research Center P.O. Box 1380, Room D2158 Houston, TX 77251-1380 LIBRARIAN Joseph L. Bingham Mantech International 2320 Mill Road Alexandria, VA 22314 LUG COORDINATOR (CORE) Dave Schmidt Management Sciences Associates 5100 Centre Avenue Pittsburgh, PA 15232 VAX-133 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX System SIG Committee List SYSTEM IMPROVEMENT REQUEST (CORE) Mark Oakley Battelle Memorial Institute 505 King Avenue Columbus, OH 43201-2669 SYMPOSIA COORDINATOR, ASSISTANT David Cossey Computer Center Union College Schenectady, NY 12308 PRE-SYMPOSIUM SEMINAR COORDINATOR HISTORIAN Jeff Jalbert J C C P.O. Box 381 Granville, OH 43023 PRE-SYMPOSIUM SEMINAR COORDINATOR (ACTING) June Baker Computer Sciences Corporation 6565 Arlington Boulevard Falls Church, VA 22046 COMMUNICATIONS, ASSISTANT David L. Wyse Professional Business Software 3680 Wyse Road Dayton, OH 45414-2539 VOLUNTEER COORDINATOR Elizabeth Bailey 222 CEB Tennessee Valley Authority Muscle Shoals, AL 35661 NEWSLETTER EDITOR Lawrence Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 VAX-134 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX System SIG Committee List SESSION NOTES EDITOR Ken Johnson Meridien Technology Corp. P.O. Box 2006 St. Louis, MO 63011 CAMPGROUND COORDINATOR Kirk Kendrick Shell Oil Company 333 Highway G, MS D-2146 Houston, TX 77082-8892 PAST CHAIR Marge Knox Computation Center University of Texas Austin, TX 78712 ADVISORS: -------- Joseph Angelico U.S. Coast Guard Detachment National Data Buoy Center NSTL Station, MS 39529-6000 Art McClinton Mitre 1820 Dolley Madison Blvd. McLean, VA 22102 Al Siegel Battelle Memorial Institute 505 King Avenue Columbus, OH 43201-2693 WORKING GROUPS: -------------- COMMERCIAL Robert Boyd GE Microelectronics Ctr. P.O. Box 13409, MS 7T3-01 Research Triangle Park, North Carolina, 27709-3049 VAX-135 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX System SIG Committee List FIELD SERVICE Dave Slater Computer Sciences Corporation 6565 Arlington Blvd. Falls Church, VA 22046 INTERNALS Carl Friedberg Seaport Systems, Inc. 165 William Street, 9th Floor New York, NY 10038-2605 LARGE SYSTEMS INTEGRATION Leslie Maltz Stevens Institute of Technology Computer Center Hoboken, NJ 07030 LIBRARY Glen Everhart 25 Sleigh Ride Road Glen Mills, PA 19342 MICROVAX Ray Kaplan Pivotal, Inc. 6892 East Dorado Court Tucson, AZ 85715-3264 (602) 886-5563 MIGRATION AND HOST DEVELOPMENT VAXINTOSH Jim Downward KMS Fusion Incorporated P.O. Box 156 D Ann Arbor, MI 48106 MULTIPROCESSOR Eugene Pal U.S. Army CAORA (ATORCATC) Fort Leavenworth, KA VAX-136 PAGESWAPPER - January 1988 - Volume 9 Number 6 VAX System SIG Committee List NETWORKS Bill Hancock P.O. Box 13557 Arlington, TX 76094-0557 REAL-TIME PROCESS CONTROL Dennis Frayne McDonnel Douglas 5301 Bolsa Aveneu Huntington Beach, CA 92646 Larry Robertson Bear Computer Systems 56512 Case Avenue North Hollywood, CA SECURITY C. Douglas Brown Sandia National Labs Division 2644 P.O. Box 5800 Albuquerque, NM 87185-5800 SYSTEM MANAGEMENT Steve Tihor 251 Mercer Street New York, NY 10012 VAXCLUSTER Thomas Linscomb Computation Center University of Texas Austin, TX 78712 PAGESWAPPER - January 1988 - Volume 9 Number 6 INPUT/OUTPUT Submission Form INPUT/OUTPUT Submission Form A SIG Information Interchange Please reprint in the next issue of the Pageswapper If this is a reply to a previous I/O, which number? ________ Caption: ______________________________________________________ Message: ______________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ Contact: Name _______________________________________________________ Address ____________________________________________________ ____________________________________________________________ ____________________________________________________________ ____________________________________________________________ Telephone ____________________________ Signature _____________________________ Date ________________ Mail this form to: Larry Kilgallen, PAGESWAPPER Editor Box 81, MIT Station, Cambridge, MA 02139-0901, USA To register for on-line submission, dial (in the United States): (617) 262-6830 and log in with the username PAGESWAPPER. PAGESWAPPER - January 1988 - Volume 9 Number 6 INPUT/OUTPUT Submission Form Tear out or photocopy reverse to submit an I/O item Larry Kilgallen, PAGESWAPPER Editor Box 81, MIT Station Cambridge, MA 02139-0901 USA PAGESWAPPER - January 1988 - Volume 9 Number 6 System Improvement Request Submission Form System Improvement Request Submission Form Page 1 of _____ ________________________________________________________________ Submittor: Firm: Address: Phone: ________________________________________________________________ How to write an SIR: Describe the capability you would like to see available on VAX systems. Be as specific as possible. Please don't assume we know how it's done on the XYZ system. Justify why the capability would be useful and give an example of its use. If you wish, suggest a possible implementation of your request. ________________________________________________________________ Abstract (Please limit to four lines): ________________________________________________________________ Description and examples (use additional pages if required) PAGESWAPPER - January 1988 - Volume 9 Number 6 System Improvement Request Submission Form Tear out or photocopy reverse to submit an SIR Mark D. Oakley Battelle Columbus Division Room 11-6-008 505 King Avenue Columbus, Ohio 43201-2369 USA