CHAPTER VAX PAGESWAPPER - October 1987 - Volume 9 Number 3 Enabling Standard VMS Image Sharing on MicroVMS VAX-3 Selectively Printing a Message-of-the-Day . . . VAX-8 Modifying the LATSYM Print Symbiont . . . . . VAX-12 Fast Definition of DCL-like Commands . . . . . VAX-15 INPUT/OUTPUT . . . . . . . . . . . . . . . . . VAX-21 VAX System SIG Committee List . . . . . . . . VAX-99 Forms at the End INPUT/OUTPUT Submission Form . . . . . . . . . VAX-104 System Improvement Request Submission Form . . VAX-106 VAX Systems SIG Fall 1987 SIR Ballot . . . . . VAX-108 PAGESWAPPER - October 1987 - Volume 9 Number 3 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 - October 1987 - Volume 9 Number 3 Enabling Standard VMS Image Sharing on MicroVMS Enabling Standard VMS Image Sharing on MicroVMS Lawrence M. Baker U. S. Geological Survey 345 Middlefield Road MS977 Menlo Park, CA 94025 We recently acquired our first MicroVAX II. It came with MicroVMS V4.5, DECnet, Fortran, and the 2-8 user license upgrade. We plan to create an environment for users on our MicroVAX II (11MB) that is similar to our existing VAX-11/750 (8MB). After restoring all the optional MicroVMS save-sets, I modified the supplied site-specific startup command procedure, Sys$Manager:SyStartup.com, to install several utilities as shareable, using the instructions found there. When I compared the MicroVMS system with the VMS system, I was surprised to see that roughly 20 or so of the remaining utilities were installed shareable on VMS, but not on MicroVMS. I had assumed that AutoGen would take notice of the ample memory and multi-user license upgrade to install the standard system utilities as shareable, but it apparently made no difference. Using the same strategy outlined in the instructions provided, I manually identified which images could be shared and removed them and re-installed them from within SyStartup.com. Each time changes were made I ran the risk of exhausting the space reserved for global pages and global sections (SysGen parameters GBLPAGES and GBLSECTIONS), which would necessitate another go-around with AutoGen. I decided that there really must be a better (i.e., automated) way to perform the installation and calculation of the SysGen parameters. After poking around and comparing the VMS system files to the MicroVMS system files, I discovered the following: 1. The initial system startup procedure, Sys$System:Startup.com, installs all standard VMS and MicroVMS images using Sys$Manager:VMSImages.dat as a script for the Install utility. VAX-3 PAGESWAPPER - October 1987 - Volume 9 Number 3 Enabling Standard VMS Image Sharing on MicroVMS 2. The AutoGen HELP text claims VMSImages.dat is created by the GENPARAMS phase, but no new versions are created when AutoGen is run on MicroVMS. 3. AutoGen will only create a new version of VMSImages.dat if either VMSIMAGES_GBLPAGES or VMSIMAGES_GBLSECTIONS is undefined. (These symbols are not located in AutoGen, nor are they documented for use in Sys$System:ModParams.dat.) 4. When AutoGen creates a new version of VMSImages.dat, it uses the value of MAXPROCESSCNT to decide whether or not certain images should be installed shareable. (Most images are always installed shareable; the conditional ones are installed shareable if MAXPROCESSCNT is greater than or equal to 25.) 5. If AutoGen creates a new version of VMSImages.dat, and either GBLPAGES or GBLSECTIONS is undefined, it supplies a value that is calculated from information contained in the new VMSImages.dat. (Comments of the form "! global_sections_required/global_pages_required" are appended to each entry.) 6. MicroVMS is supplied with a special version of VMSImages.dat and a version of Sys$System:VMSParams.dat containing definitions for VMSIMAGES_GBLPAGES, VMSIMAGES_GBLSECTIONS, GBLPAGES, and GBLSECTIONS. As a result, AutoGen bypasses both the creation of a new VMSImages.dat and the calculation of GBLPAGES and GBLSECTIONS on MicroVMS. I inspected the value of MAXPROCESSCNT in my MicroVMS system and found it was over 25, so all that had to be done was to figure out a way to "undefine" an AutoGen parameter. Since I preferred to stay within the bounds of modifying the system using only ModParams.dat for the AutoGen phase and SyStartup.com for the boot phase, I looked for a method that could be implemented within ModParams.dat. During the GETDATA phase, AutoGen processes several parameter files to generate Sys$System:Params.dat. These include VMSParams.dat and ModParams.dat, which, if they exist, are appended to the end of Params.dat, in that order. Thus, it might be possible to undo something in Params.dat that came from VMSParams.dat with one or more lines that came from VAX-4 PAGESWAPPER - October 1987 - Volume 9 Number 3 Enabling Standard VMS Image Sharing on MicroVMS ModParams.dat. Each line in Params.dat is eventually transformed by AutoGen into a DCL one-liner that is "executed" roughly according to the following rules: 1. A "symbol name" is extracted by trimming leading and trailing blanks from the portion of the line preceeding an equals sign, or the full line, if there is no equals sign. 2. If the symbol name begins with "!", the line is ignored. 3. Symbol names beginning with "VMS$" or "ADD_" are handled separately. 4. If the symbol name contains embedded blanks, the entire line is immediately passed to DCL. 5. Eventually, any other line of the form "symbol = expression" is passed to DCL. 6. Anything else will probably cause AutoGen to fail. The simplest solution is to add two lines to ModParams.dat that enable the generation of a new VMSImages.dat: Delete/Symbol VMSIMAGES_GBLPAGES Delete/Symbol VMSIMAGES_GBLSECTIONS and two lines that enable the automatic calculation of GBLPAGES and GBLSECTIONS: Delete/Symbol GBLPAGES Delete/Symbol GBLSECTIONS The "symbol name" in each case is the whole line, since there are no equals signs. Each line is then passed directly as is to DCL, since the symbol name contains an embedded blank. This will yield a customized ModParams.dat file that will work, but only on a MicroVMS system, since these symbols are defined in VMSParams.dat, which is not included with VMS. I like to keep both my MicroVMS and VMS systems as similar as possible, which is easier if I can use the same ModParams.dat file for VAX-5 PAGESWAPPER - October 1987 - Volume 9 Number 3 Enabling Standard VMS Image Sharing on MicroVMS both. Also, AutoGen could fail if DEC later removes one or more of these definitions from VMSParams.dat (e.g., if the distinction goes away between MicroVMS and VMS), so I added a few safety features. The final version I use is given below (remember: no continuation lines are allowed -- only one-liners). It specifically checks for the symbols before deleting them, and only deletes GBLPAGES and GBLSECTIONS if it deleted VMSIMAGES_GBLPAGES and VMSIMAGES_GBLSECTIONS, respectively. On MicroVMS, the entry in Sys$Manager:VMSImages.dat for Sys$Share:mthrtl must be removed, since the correct version of mthrtl depends on the processor type. (Sys$System:UVStartup.com installs the proper version of mthrtl as part of the standard MicroVMS boot procedure.) Instructions are provided in Sys$Manager:SyStartup.com to install mthrtl as shareable. Only one image is included in VMSImages.dat that is not a part of the MicroVMS distribution kit: Sys$System:ShwClstr.exe. (I assume this is part of the DCL Show command that deals with clusters.) You can either ignore the error messages at startup or remove the entry in VMSImages.dat. The original MicroVMS version of VMSIimages.dat also contains entries for several shareable libraries from Sys$Message (adamsg, scnmsg, and vaxcmsg) that are apparently used by layered products. You can include these by comparing the original VMSImages.dat with the new VMSImages.dat and manually inserting entries for the libraries you need. As an aside, I also found that AutoGen sets the default size of large request packets (LRPSIZE) to 1504 if it finds an Ethernet controller, on the assumption that the DECnet buffer size will also be 1504. If this is not the case (as with our DECnet), you can reduce the LRPSIZE parameter to the actual DECnet buffer size. VAX-6 PAGESWAPPER - October 1987 - Volume 9 Number 3 Enabling Standard VMS Image Sharing on MicroVMS Disclaimer Although this program has been tested by the Geological Survey, United States Department of the Interior, no warranty, expressed or implied, is made by the Geological Survey as to the accuracy and functioning of the program and related program material nor shall the fact of distribution constitute any such warranty, and no responsibility is assumed by the Geological Survey in connection therewith. Sys$System:ModParams.dat ! ! Install everything /Shared on a MicroVAX (undoes damage from ! VMSParams.dat). (Must comment out the line containing ! Sys$System:ShwClstr.exe in Sys$Manager:VMSImages.dat to avoid ! the error message during startup.) ! ! Force the creation of a new version of VMSImages.dat ! del_vmsgblpages = F$TYPE(vmsimages_gblpages) .nes. "" del_vmsgblsections = F$TYPE(vmsimages_gblsections) .nes. "" If (del_vmsgblpages) Then Delete/Symbol vmsimages_gblpages If (del_vmsgblsections) Then Delete/Symbol vmsimages_gblsections ! ! Force calculation of these parameters from VMSImages.dat ! del_gblpages = F$TYPE(gblpages) .nes. "" del_gblsections = F$TYPE(gblsections) .nes. "" del_gblpages = del_gblpages .and. del_vmsgblpages del_gblsections = del_gblsections .and. del_vmsgblsections If (del_gblpages) Then $ Delete/Symbol gblpages If (del_gblsections) Then $ Delete/Symbol gblsections ! ! Cleanup ! Delete/Symbol del_vmsgblpages Delete/Symbol del_vmsgblsections Delete/Symbol del_gblpages Delete/Symbol del_gblsections ! LRPSIZE = 576 ! Since our DECnet doesn't use 1504 VAX-7 PAGESWAPPER - October 1987 - Volume 9 Number 3 Selectively Printing a Message-of-the-Day Selectively Printing a Message-of-the-Day Lawrence M. Baker U. S. Geological Survey 345 Middlefield Road MS977 Menlo Park, CA 94025 It is important that the system manager have a way to inform users of system changes or other items of interest. This is typically accomplished by printing system messages on the user's terminal immediately after a successful login. Since most of the time the messages do not change, it is somewhat annoying to sit through the same system messages every time you log into a computer system. Unfortunately, this leads to a situation where the messages are ignored most of the time, thus no longer serving their intended purpose. The attached command procedure, MOTD.com, is a simple attempt to reduce the unnecessary printing of the same messages every time a user logs in, while also making sure that new messages are seen as soon as possible. When called from the system-wide login command procedure, Sys$SyLogin, MOTD.com prints a message-of-the-day file, Sys$Manager:MOTD.txt, when a user logs in for the first time of a day, or if the contents of MOTD.txt have changed since the last login. (If MOTD.txt does not exist, the procedure simply exits.) A file is maintained in the user's login directory, Sys$Login:MOTD.dat, which contains the date and time of the last login. MOTD.txt is printed under the following conditions: 1. MOTD.dat does not exist. 2. The current login date is greater than the last login date. 3. The login dates agree, and the revision date/time of MOTD.txt is greater than the last login date/time. Regardless of whether MOTD.txt is printed, MOTD.dat is updated with the current login date and time. VAX-8 PAGESWAPPER - October 1987 - Volume 9 Number 3 Selectively Printing a Message-of-the-Day MOTD.txt is actually a command procedure, not a text file. It contains text, sandwiched between a DCL Type command and Exit command (see below). This allows MOTD.com to define the symbol MOTD to be @Sys$Manager:MOTD.com for printing the message-of-the-day on demand any time. It also allows for the file to contain a history of messages (between the DCL Exit command and the actual end-of-file), which can always be viewed by simply typing or editing MOTD.txt. (This is particularly useful as a reference when the system manager prepares activity reports.) No distinction is made in MOTD.com between an INTERACTIVE, BATCH, or NETWORK login --- it is assumed that the outer-level command procedure will make that choice. (You can of course modify MOTD.com to suit your own preferences. On our systems, MOTD.com is only called for INTERACTIVE logins.) Disclaimer Although this program has been tested by the Geological Survey, United States Department of the Interior, no warranty, expressed or implied, is made by the Geological Survey as to the accuracy and functioning of the program and related program material nor shall the fact of distribution constitute any such warranty, and no responsibility is assumed by the Geological Survey in connection therewith. Sys$Manager:MOTD.com $ $! MOTD.com - Print the system message-of-the-day once a day (or $! the first time after it has changed). $ $! Version 2.0 26-Jun-1987 $ $! This command file is called from the system-wide login com- $! mand file, Sys$SyLogin. It tries to locate a file in the $! user's login directory that contains the last time the user $! logged on (MOTD.dat). If MOTD.dat does not exist, or if this $! is the first login of the day, or if the message-of-the-day $! file (Sys$Manager:MOTD.txt) has been changed since the last $! login, then the message-of-the-day is printed. In any case, $! the login time is saved in MOTD.dat. $ $! Sys$Manager:MOTD.txt is really a command procedure, not a VAX-9 PAGESWAPPER - October 1987 - Volume 9 Number 3 Selectively Printing a Message-of-the-Day $! text file. This is so that a historical record is maintain- $! ed in the same file for future reference. Typically, the $! file looks like this: $ $! $ Type Sys$Input $! Date of message Message text... $! $ Exit $! Old messages... $! [EOF] $ $! To view the current message-of-the-day, type "MOTD". To view $! all messages in the file, type or edit Sys$Manager:MOTD.txt. $ $! Modification History: $ $! 26-Jun-1987 L. M. Baker Use F$CvTime for date/time $! comparisons. $! Use RDT instead of CDT for $! MOTD.txt date/time. $! Use F$GetJPI instead of F$Time $! for login time. $ $! Anyone know how to get at last interactive and batch login $! date/time maintained by the system (printed by loginout)? $ $ MOTD :== @Sys$Manager:MOTD.txt $ $! Disable this feature if Sys$Manager:MOTD.txt does not exist $ $ motd_file = F$Search("Sys$Manager:MOTD.txt") $ If (motd_file .eqs. "") Then $ Goto EXIT $ $! If there is no last login time file, display the message $ $ If (F$Search("Sys$Login:MOTD.dat;1") .eqs. "") Then - $ Goto PRINT_TEXT $ $! If the current date is later than the last login date, $! display the message $ $ login_time = F$GetJPI("","LOGINTIM") $ last_time = "" $ Open/Read/Error=OPEN_ERROR save_file Sys$Login:MOTD.dat;1 $ Read/Error=READ_ERROR save_file last_time $READ_ERROR: $ Close save_file $OPEN_ERROR: VAX-10 PAGESWAPPER - October 1987 - Volume 9 Number 3 Selectively Printing a Message-of-the-Day $ Delete Sys$Login:MOTD.dat;* $ If (last_time .eqs. "") Then $ Goto PRINT_TEXT $ If (F$CvTime(login_time,"COMPARISON","DATE") .gts. - F$CvTime(last_time, "COMPARISON","DATE")) Then - $ Goto PRINT_TEXT $ $! The dates are the same. If the message file has changed $! since the last login, display the message $ $ motd_time = F$File_Attributes(motd_file,"RDT") $ If (F$CvTime(motd_time,"COMPARISON","DATETIME") .les. - F$CvTime(last_time,"COMPARISON","DATETIME")) Then - $ Goto SAVE_TIME $ $PRINT_TEXT: $ MOTD $ $SAVE_TIME: $ Open/Write/Error=EXIT save_file Sys$Login:MOTD.dat;1 $ Write/Error=WRITE_ERROR save_file login_time $WRITE_ERROR: $ Close save_file $ $EXIT: Sys$Manager:MOTD.txt $ $! MOTD.txt - Cumulative system message-of-the-day file. $ $! This is a template message-of-the-day file for use with the $! MOTD.com command procedure. Place the current messages be- $! tween the DCL Type and Exit commands (no $'s in column 1, $! please). Place old messages between the DCL Exit command and $! the end-of-file. $ $ Type Sys$Input $ Exit VAX-11 PAGESWAPPER - October 1987 - Volume 9 Number 3 Modifying the LATSYM Print Symbiont Modifying the LATSYM Print Symbiont Robert W. Shew, CS '89 Waters Computing Center Rose-Hulman Institute of Technology 5500 Wabash Avenue RHIT Box 19 Terre Haute, IN 47803-3999 (812) 877-1511 Ext. 239 DEC supplies with LATPlus/VMS a print symbiont modification routine (PSM) called LATSYM.EXE for remote printers on DECservers. The use of this image prohibits the implementation of any other PSM routines that an installation may have developed. We decided that we needed to use our PSM routines and the DEC routines. The best solution was to replace the standard symbiont routines as we had before LATSYM was available, and then somehow activate the LATSYM image. The LATSYM image was activated with the system services $ImgAct_S and $ImgFix_S. These system services merge our image with the LATSYM image, allowing us to use both sets of PSM routines. However, LATSYM replaces the routines Job_Completion and Output. Therefore, you cannot substitute these two routines if you want to use LATSYM, because the substitution will have no effect. Here is the main PSM routine written in VAX Macro: .Library "SYS$LIBRARY:LIB.MLB" $PSMDef ; PSM definitions $IACDef ; Image Activation definitions $IHADef ; Image Header Address ; definitions $IHDDef ; Image Header Descriptor ; definitions Flag_Code: .Long PSM$K_File_Flag ; PSM$Replace code Header_Buff: .BlkB 512 ; Buffer for image header Image_Desc: .AsciD "SYS$SYSTEM:LATSYM" Def_Filename: .AsciD "[].EXE" VAX-12 PAGESWAPPER - October 1987 - Volume 9 Number 3 Modifying the LATSYM Print Symbiont ImgAct_Ctrl: .Long IAC$M_Merge!IAC$M_Expreg ; Merge and map into next ; available spot in P0 space .Entry RHITPSM, ^M<> PushAW Print_Flag ; New flag page routine ;*** Print_Flag is the address of the entry mask for our ;*** PSM replacement routine. This is where you would put ;*** the address of the entry mask for your routine. PushAL Flag_Code ; PSM code for flag page ;*** Insert the proper code for the routine that you wish ;*** to replace. Calls #2, G^PSM$Replace ; Replace theirs with ours BlBS R0, 10$ ; Check for error Ret ; Return if error occurred 10$: $ImgAct_S Name=Image_Desc, - ; Merge the images Dflnam=Def_Filename, - ; in P0 space Hdrbuf=Header_Buff, - Imgctl=ImgAct_Ctrl BlBS R0, 20$ ; Check for error Ret ; Return if error occurred 20$: $ImgFix_S ; Perform address relocation BlBS R0, 30$ ; Check for error Ret ; Return if error occurred 30$: MovL Header_Buff, R1 ; Address of IHD MovZWL IHD$W_Activoff(R1), - R1 ; Offset to activation data AddL Header_Buff, R1 ; Address of transfer vector CallS #0, @IHA$L_TfrAdr2(R1) ; Call the new routine ; Want SECOND transfer ; address Ret ; Return to caller .End RHITPSM ; Transfer Address VAX-13 PAGESWAPPER - October 1987 - Volume 9 Number 3 Modifying the LATSYM Print Symbiont When the RHITPSM image is executed, all the necessary sections of the shareable Run-Time Library routines are mapped into the image, and the first set of PSM routines replaced. Then the LATSYM image is mapped in and merged with the RHITPSM image. The RHITPSM routine then calls the LATSYM routine, transferring control to it. The fourth argument to the $ImgAct_S system service holds the control flags for the routine. The IAC$M_Merge flag specifies that the images are to be merged. The IAC$M_Expreg flag specifies that the second image will be mapped into the next available slot in P0 space. A more detailed description of image activation may be found in Chapter 21, _V_A_X_/_V_M_S___I_n_t_e_r_n_a_l_s___a_n_d___D_a_t_a___S_t_r_u_c_t_u_r_e_s, Lawrence J. Kenah and Simon F. Bate. The data structure of the image file and image header is described in Appendix B, _V_A_X_/_V_M_S___I_n_t_e_r_n_a_l_s___a_n_d___D_a_t_a___S_t_r_u_c_t_u_r_e_s. Using the DCL command ANALYZE/IMAGE SYS$SYSTEM:LATSYM showed that the second transfer address in the image was the address of the main routine. NOTE: These system services are not in the VAX/VMS documentation because they are not meant to be called by users. Because of this, a future version of VAX/VMS may cause a problem with this method and care should be taken. If you want to obtain a copy of the source, you may send a tape to the above address. If you have any questions, I can be reached at the phone number listed above. VAX-14 PAGESWAPPER - October 1987 - Volume 9 Number 3 Fast Definition of DCL-like Commands Fast Definition of DCL-like Commands Frank J. Nagy Fermi National Accelerator Laboratory P. O. Box 500 Mail Stop 220 Batavia, IL 60510 The Command Definition Utility is very helpful in writing new utilities and software tools. The CLI$ routines make it very easy to write programs which are invoked as DCL-like commands. The problem is that SET COMMAND is relatively slow; this precludes using SET COMMAND to add the new commands to your DCLTABLES on each login. In addition, as each user does this a private copy of the shared DCLTABLES section is created which impacts the utilization of memory and paging file. On the other hand, few system managers are prepared, at this time, to insert new commands into the system-wide DCLTABLES in a permanent fashion since such commands have a frightening ability to disappear during system upgrades. Conversely, the foreign command definition facility of DCL allows a new command to be defined very quickly with a single line in a login file: name :== $disk:[directory]image.exe The problem with this approach is that the burden of decoding the command line is back on the programmer who cannot, apparently, appeal to DCL for assistance. What is presented below is a description of how to merge these two methods and end up with new DCL-like commands which require minimal time to define but which ease the programmer's burden by making use of the CLI$ routines to parse the command line. Tool Using SET COMMAND We begin with the definition module (.CLD file) for a simple tool. This module defines the tool as a DCL command. In order to use the tool, SET COMMAND must be done to enter the command VAX-15 PAGESWAPPER - October 1987 - Volume 9 Number 3 Fast Definition of DCL-like Commands in the DCL tables for each login. Define Verb DOESNOTHING Image disk:[directory]DoesNotAThing Parameter P1, Prompt="To Not Do", Value(Required) Qualifier LOG Qualifier NONE This command definition file is then used with SET COMMAND to define the command in the DCLTABLES. $ SET COMMAND disk:[directory]DOESNOTHING.CLD Next we have a fragment of the FORTRAN program which implements this command. Program NotAThing C Integer*4 Cli$Get_Value, Cli$Present, sts Character*63 Parameter Logical*1 Logging C C Check for /LOG qualifier and set flag C Logging = Cli$Present( 'LOG') C C Get parameter value C sts = Cli$Get_Value( 'P1', Parameter) IF (.NOT. sts) CALL Lib$Signal( %VAL(sts)) C C If /NONE qualifier is not present, then process parameter C IF (.NOT. Cli$Present( 'NONE')) THEN CALL Process( Parameter, Logging) ENDIF C END The program uses the CLI$ routines to check for the presence of qualifiers and to access values of parameters and qualifiers. In this way, using the Command Definition Utility simplifies the creation of new tools using DCL-like interfaces. VAX-16 PAGESWAPPER - October 1987 - Volume 9 Number 3 Fast Definition of DCL-like Commands Same Tool as Foreign DCL Command Now we take the command definition module and code fragment from above and modify them to create a tool which we can define as a foreign command. This permits rapid definition of the tool's command verb (particularly important when the system is rich in software tools) without sacrificing the ability to use the DCL command parser. This is done by modifying the command definition module to remove the "Image" declaration line and add an initial line to specify a module name and external symbol for the command tables. Module DN_Tables Ident "V1.0" Define Verb DOESNOTHING Parameter P1, Prompt="To Not Do", Value(Required) Qualifier LOG Qualifier NONE This command definition file is then processed with SET COMMAND to create an object module with the command definition tables as: $ SET COMMAND/OBJECT=DNTABLES DOESNOTHING.CLD Next we modify the FORTRAN program with some additional declarations. Note, in particular, the declaration of the external symbol DN_Tables. This declaration will be used to connect the command definition tables in the above object module with the DCL command parser. Program NotAThing C External DN_Tables Integer*4 Lib$Get_Foreign, Cli$Dcl_Parse, Lib$Get_Input Character*255 CmdLine Integer*2 CL_Len Integer*4 Cli$Get_Value, Cli$Present, sts Character*63 Parameter Logical*1 Logging C C Get the Foreign Command line, tack the verb onto the front and VAX-17 PAGESWAPPER - October 1987 - Volume 9 Number 3 Fast Definition of DCL-like Commands C invoke the DCL command processor on the result. C sts = Lib$Get_Foreign( CmdLine,, CL_Len) IF (.NOT. sts) CALL Lib$Signal( %VAL(sts)) IF (CL_Len .GT. 0) THEN sts = Cli$Dcl_Parse( 'DOESNOTHING '//CmdLine(1:CL_Len), 1 DN_Tables, Lib$Get_Input) ELSE sts = Cli$Dcl_Parse( 'DOESNOTHING ', 1 DN_Tables, Lib$Get_Input) ENDIF IF (.NOT. sts) CALL Exit C C Check for /LOG qualifier and set flag C Logging = Cli$Present( 'LOG') C C Get parameter value C sts = Cli$Get_Value( 'P1', Parameter) IF (.NOT. sts) CALL Lib$Signal( %VAL(sts)) C C If /NONE qualifier is not present, then process parameter C IF (.NOT. Cli$Present( 'NONE')) THEN CALL Process( Parameter, Logging) ENDIF C END The other change adds a few lines of code at the very start of the program. This code calls Lib$Get_Foreign to get the foreign command line (without the command verb). The string "DOESNOTHING" is then prepended to the results from Lib$Get_Foreign; this forms a complete command line for the DCL command processor. Again, this prepended command verb must match the "Define Verb" declaration in the command definition tables module. Also, a separating space must provided above between the command verb and the rest of the command line gotten from Lib$Get_Foreign as show in the example. VAX-18 PAGESWAPPER - October 1987 - Volume 9 Number 3 Fast Definition of DCL-like Commands The DCL command processor is then invoked by the call to Cli$Dcl_Parse, passing to it the command line to be parsed, the location of the command definition tables (via the external DN_Tables) and a routine to call to prompt for missing parameters, in this case Lib$Get_Input. The program is then linked with the command definition table object module: $ LINK DOESNOTHING,DNTABLES thus the command definition becomes part of the program image itself. Once the foreign command line has been read and turned into a pseudo-DCL command and parsed by the DCL command parser, all the internal DCL tables are setup for use by the other CLI$ routines. The remainder of the program with its calls to Cli$Present and Cli$Get_Value is unchanged between the first example and this one. The difference is that the command definitions for tools created using the method just described do not need to be entered into the permanent, system-wide DCL tables, do not slow user logins to a crawl with multiple SET COMMAND calls and lessen the impact on the system memory of multiple, individual copies of the DCLTABLES section (rather than all the processes sharing the single, installed copy). The command is quickly defined as a foreign DCL command and the command definition tables used to decode the command line are part of the program image. Finis The ability to use the DCL command parser for user-written utilities is a major boon to both programmers (lessens development time) and users (a uniform syntax for commands). Creative use of this facility along with the DCL foreign command definitions can allow a system to have a rich and varied set of tools with minimal impact upon usage of system memory (DCLTABLES remains shared) and user login time. The techniques discussed here can be applied to tools written in languages other than FORTRAN. In particular, several tools written in VAX C have been developed at Fermilab. In addition, with just a few hours of work, programs from VAX SIG tapes which used a SET COMMAND to define the command in the DCL tables have VAX-19 PAGESWAPPER - October 1987 - Volume 9 Number 3 Fast Definition of DCL-like Commands been modified to use this foreign command technique. VAX-20 PAGESWAPPER - October 1987 - Volume 9 Number 3 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 585.14 Anyone use defrag programs? 14 of 15 "John Burnet" 20 lines 22-AUG-1987 13:44 -< Rabbit-7 (Dopter) experiences wanted >- ---------------------------------------------------------------- Speaking of disk defrag programs... We're looking at Raxco's Rabbit-7 product, also known as DOPTER. We've got a demo copy and are thinking of buying the program. The price isn't too bad, the documentation is fairly good, and we've used the demo copy to compress several of our disks a few times with good results. Question: Does anyone have any comments or experiences to share regarding this product, or other similar (specific brand-name) products? It would help us to have some accounts of first-hand experiences (pros and cons, especially cons... horror stories are invited), since all we have to go on now is the vendor's literature and reviews from Digital Review, Digital News, etc. John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 VAX-21 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 585.15 Anyone use defrag programs? 15 of 15 "Lee K. Gleason" 25 lines 24-AUG-1987 22:53 -< Wascally Wabbits >- ---------------------------------------------------------------- I've been using this product on eight RA-81s in a three VAX clusters, on a weekly basis, for approx. 4 months now, and it hasn't trashed anything yet. However, I haven't worked up the nerve to try it on the system disk (I mean, that's were all MY files are). Nor have I the courage to use it when the disks are being actively used (I use a command file, run late at night). I take a pretty conservative approach with internal type stuff that didn't come from DEC. It does a pretty good (but not perfect) job of compressing a disk back to usable shape. It's command interface suffers from it's heritage - it has a real RSTS/e sort of feel to it - you'll never mistake it for a DEC VMS product. On the whole, I've been pleasantly surprised, since I really didn't want to use a third party product for something like this, and expected all sorts of problems that just have not occurred. If you want to use a third party defragmenter, this is not a bad choice. Lee K. Gleason Control-G Consultants 2416 Branard #D Houston TX 77098 713/528-1859 ================================================================ Note 599.12 Any PSI users out there? 12 of 12 "Offline Submission" 13 lines 30-AUG-1987 20:55 -< We are PSI users >- ---------------------------------------------------------------- We use PSI for PSI mail, DECnet data links and X.29 virtual terminals. Starting PSI was not easy, but now it works pretty well. Here in France, many VAXes are connected to the X.25 network (named TRANSPAC) and use PSI, but of course costs vary from one country to another. Nicole Bonis Universite de Compiegne VAX-22 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT 60206 Compiegne France Date: August 13, 1987 ================================================================ Note 608.2 VMS 4.5A installation - LAVC 2 of 2 "John Osudar" 18 lines 25-AUG-1987 16:25 -< 4.5C still is, but... >- ---------------------------------------------------------------- We, too, found installing LAVc a pain in the neck. What we did was to install V4.5A, V4.5C, security update, LAVc key, etc. on a clean disk; then get rid of all files in that system that we knew we wanted to replace with files from our pre-LAVc V4.5 system (e.g. HELPLIB.HLB, DCLTABLES.EXE, etc.) Then we took (a copy of) our V4.5 system disk, deleted all files from it that would be replaced by the V4.5C files, and merged the V4.5C files into it. This system booted successfully, so we ran BOOT_CONFIG on it and rebooted as a boot member in a (one-node) LAVc. Then we moved the bare minimum files from the [SYS0] tree into [V4COMMON] (things like DCLTABLES) and did the SATELLITE_CONFIG. Once we got this right (and it took a while to figure out that DCLTABLES was missing!) our new VAXstation 2000 booted successfully as a satellite node. Finally, I spent an afternoon renaming other files from the [SYS0] tree into [V4COMMON], and did one last reboot to bring things together. Total (useful) labor, about 12 hours; we've been up for half a week with no problems, and we reinstalled NO layered products (and we run a ton of them, both DEC and non-DEC.) John Osudar Argonne National Laboratory 9700 S. Cass Ave. Bldg. 205 A-051 Argonne, IL 60439-4837 (312) 972-7505 VAX-23 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 644.19 Wanted: MAIL improvements 19 of 20 "Robert Blunt" 4 lines 4-AUG-1987 11:51 -< All for a Callable interface >- ---------------------------------------------------------------- All of those additions and enhancements would be nice, but just having a callable interface would be sufficient for most issues. Standardization may be the answer to the problems related to compatibility between implementations. Robert Blunt INTERGRAPH CORPORATION 1 Madison Industrial Park Huntsville, Al. 35807 Mail Stop CR1100 ================================================================ Note 644.20 Wanted: MAIL improvements 20 of 20 "John Osudar" 29 lines 4-AUG-1987 12:52 -< yes, but... >- ---------------------------------------------------------------- Callable interfaces are nice. I vote for callable interfaces to everything (well, maybe not OPCCRASH, but...) Standardization is nice, too. Except that in many cases (particularly where users have to implement system-level software), standardization is only something that you wish for. User-written software usually gets "standardized" in one of two ways: (a) one person writes a version that's so much better (or, sometimes, gets it done so much sooner) than others, that everyone else adopts that version as the "standard"; or (b) a committee of users (e.g. a Sig working group) gets together and decides what the standard features should be, and someone implements that. The BEST way to "standardize" system-level software is to have DEC implement it. Seems to me that if their callable interfaces make it so easy for users to "roll their own", it should be equally easy for DEC to give us a product with the most-desired enhancements already built in. There's one other problem with relying solely on the "callability" of VMS software: you can only do as much as the underlying protocols allow you to do. For example, many VMS mail interfaces based on the Mail-11 protocol have been implemented over the years, but none of them can support return receipts VAX-24 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT without using one kludge or another, because Mail-11 has no provision for return receipts. If DEC provides a callable interface but does no work on enhancing underlying protocols, you still won't be able to do what you want -- or at best you'll end up with a bunch of "standard" kludges. If DEC DOES enhance underlying protocols, then 90% of the work required to implement those enhancements in the product will be done -- so why not do the other 10%? John Osudar Argonne National Laboratory 9700 S. Cass Ave. Bldg. 205 A-051 Argonne, IL 60439-4837 (312) 972-7505 ================================================================ Note 656.3 Replacement of BNT 3 of 4 "Gus Altobello" 25 lines 31-JUL-1987 00:17 -< More worries, paranoia, etc. >- ---------------------------------------------------------------- Re: .2, "T" as in "tape"... Never heard of that one, and the 8530 we got had no tape drives attached. Wonder what the "A" in "DEBNA" (the replacement device) stands for? Now that the DEBNA is installed, I'm getting reports that throughput tests are turning up strange results. If the DTS/DTR test is run to our 8600 node (which has a DEUNA) then all seems to be ok. But adding simultaneous DTS/DTR processes to other nodes to increase the throughput seems to make the circuit bounce. Are there any other DEBNA users out there, satisfied, unsatisfied, or undecided, who could let me know what to expect from this device? My experiences with the DEBNT, now replaced by the DEBNA, are making me a bit unsure about migrating my Development machine, and its 100 users, from our poor overworked 8600 to the 8530 cluster. Gus Altobello PO Box 11274 Hauppauge, NY 11788 516/435-7036 VAX-25 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 656.4 Replacement of BNT 4 of 4 "Brian Tillman, Lear Siegler, Inc." 1 line 31-JUL-1987 14:16 ---------------------------------------------------------------- DEBNA: Digital Ethernet BI Network Adaptor Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 ================================================================ Note 670.4 Set Watch 4 of 4 "Lee K. Gleason" 44 lines 10-AUG-1987 00:39 -< The Way Things Work >- ---------------------------------------------------------------- I had stumbled across the code for the SET WATCH command in the XQP micro-fiche, and experimented with it for quite a while before realizing there was a DCL command to activate it. Here's some info on what makes SET WATCH work internally. A little study of the code showed that the trace functions are controlled by the state of six bits of a byte in each processes P1 space. At the end of any XQP function, the appropriate bit in the byte is tested, and an AST is used to print out a message on the screen. The byte in question is the low byte at PIO$GW_DFPROT. At first, this was a little confusing to me, since this word in the past has been documented as the location of the default protection for newly created files. Some tests with SET PROTECTION/DEFAULT and ANALYZE/SYS revealed that this is no longer the case, and apparently the word (well, the low byte of the word) has been taken over for controlling this trace function. My guess is that it was convenient for the developer to use it, since it has a globally defined symbolic address, and it is (well, it used to be) a long time between VMS executive re-linkings. Bit 0 Major functions Bit 1 Control functions Bit 2 Read or Write file attributes Bit 3 Directory operations VAX-26 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Bit 4 Display entire FIB on every operation Bit 5 Disk Quota functions Bit 6 ...No apparent effect... Bit 7 ...Likewise... Anyway, it is no problem for any program that can get into EXEC mode to turn these bits on and off, so you can start and stop these traces in mid flight, without exiting to DCL to executing a SET WATCH command. Clearing the byte before you do anything interesting, then restoring it after, is a good way to keep the unwashed from seeing how your Evil Magic works. Lee K. Gleason Control-G Consultants 2416 Branard #D Houston TX 77098 713/528-1859 ================================================================ Note 680.3 Looking for DEC PCFS Experiences 3 of 7 "Bill Mayhew" 9 lines 10-AUG-1987 17:22 -< DECnet/DOS performance troubles? >- ---------------------------------------------------------------- I spoke with someone today who has a client using DECnet/DOS who is having what seem, to him and me, to be fairly severe performance problems. It takes them two minutes to load 1-2-3 into a PC over the network, for instance. At this time I know nothing about the specifics (like what kind of VAX host, what kind of PC, what kind of other load on the host) but I am wondering if anybody else has had similar experiences. There is some finger-pointing going on at the moment and a desire to end it. Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 VAX-27 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 680.4 Looking for DEC PCFS Experiences 4 of 7 "John Burnet" 20 lines 19-AUG-1987 12:26 -< Async DDCMP performance >- ---------------------------------------------------------------- For what it's worth... I recently heard a tape of a Fall '86 DECUS session on async DECnet (which is, I presume, what DECnet-DOS uses, as opposed to PCFS, which uses Ethernet). In the Q & A period at the end, someone asked the speaker (who was from DEC) to quantify async DECnet's performance as a ratio of observed data bits/sec. to async line speed--e.g., during a file transfer. The speaker became rather flustered, gave an evasive answer, and refused to provide any numbers. Wonder why... (he said innocently) We've found async DDCMP performance to be okay (nothing great) over 19.2K-baud hardwired lines, but it's pretty poor over a 1200-baud phone connection. Btw, is there anyone reading this who *is* able and willing to tell us what that magic async-DECnet efficiency ratio is? Is it really bad enough to be cause for embarrassment to DEC? John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 ================================================================ Note 680.5 Looking for DEC PCFS Experiences 5 of 7 "Brian Tillman, Lear Siegler, Inc." 14 lines 20-AUG-1987 08:57 -< DECnet-DOS is not media dependent >- ---------------------------------------------------------------- RE: < Note 680.4 by NODE::US124053 "John Burnet" > -< Async DDCMP performance >- VAX-28 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT I recently heard a tape of a Fall '86 DECUS session on async DECnet (which is, I presume, what DECnet-DOS uses, as opposed to PCFS, which uses Ethernet). DECnet-DOS will use whatever the physical medium is, just like DECnet-VMS. We use DECnet-DOS on Ethernet for almost all our PCs. Unfortunately, at the time we put out PCs on the Ethernet, the DEC Ethernet controller for PCs was unavailable, so we used 3-Com. The problem with that it that DECnet-DOS can't use the DMA feature of those boards, so load is heavier on the PCs and the speed suffers (a lot). Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 ================================================================ Note 702.5 Planning an Ethernet 5 of 5 "Marc A. Shannon" 29 lines 8-AUG-1987 13:23 -< You might want to check out CMU's TCP/IP too >- ---------------------------------------------------------------- Something you might want to check into for a (very) cheap TCP/IP is CMU's TCP/IP (originally from Tektronix, but now CMU is handling all distributions). This TCP/IP is available for only the cost of distribution (I've heard $100, but I'm not sure). With this, plus maybe PMDF (from Harvey Mudd College) interfacing with it, you'd have a very nice TCP/IP system for little money. The sites on campus at CMU which were running Wollongong (albeit version 2, not 3), have switched to the CMU TCP/IP. For more information on the CMU TCP/IP, send inquiries to: CMU TCP/IP Distribution 4910 Forbes Avenue Pittsburgh, PA 15213 Information on PMDF can be obtained from: Ned Freed Harvey Mudd College Claremont, CA VAX-29 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT I use both of the above on a MicroVAX-I and am very satisfied with their performance. Marc A. Shannon CMU P. O. Box 257 Pittsburgh, PA 15213 ================================================================ Note 707.3 Disk Serving for MicroVax? 3 of 6 "Bob Hassinger" 7 lines 31-JUL-1987 12:10 -< Yes, but that is just the point... >- ---------------------------------------------------------------- Yes, I recall, but that is just the point! That is one of the statements I mentioned hearing at various times. What some developer said at some LUG meeting does not hold much water when you have trouble. How can anyone actually commit to the thousands of dollars and the whole strategic approach for their computer facilities without a formal statement from DEC committing to make it work? Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 707.4 Disk Serving for MicroVax? 4 of 6 "Frank J. Nagy" 13 lines 1-AUG-1987 10:22 -< Can almost guarantee it would work... >- ---------------------------------------------------------------- 1. Yes, you will need LAVC to serve the disks from your 780 to the MicroVAXes. 2. I cannot verify that non-RA disks will be served by an LAVC since my LAVC's only have RA and RD disks on them. However, the MSCP Server used by the LAVC is IDENTICAL (repeat IDENTICAL, SAME EXACT, etc.) module used on the CI clusters and there I can verify that Massbus disks get served up just fine and dandy. Remember, an LAVC is, software-wise, essentially identical to a CI cluster except for the underlying (CI, HSCs, CI780s) hardware. We have found that an LAVC can do whatever a VAX-30 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT CI cluster can do. Therefore, I would be vastly surprised that Massbus disks would not be served in an LAVC! Frank J. Nagy Fermilab PO Box 500 MS/220 Batavia, IL 60510 (312)840-4935 ================================================================ Note 707.5 Disk Serving for MicroVax? 5 of 6 "DALE E. COY" 42 lines 21-AUG-1987 18:02 -< LAVC Hints >- ---------------------------------------------------------------- { How can anyone actually commit to the thousands of dollars and the whole strategic approach for their computer facilities without a formal statement from DEC committing to make it work?} We all do that all the time ("Will be fixed in a future release..."). Seriously, that's what this Conference is all about. I agree with the previous comment that it HAS to work if it's MSCP. A satellite's RDs are MSCPd to the LAVC. Of course, Bob, I haven't the hardware to demonstrate it, and as you point out it's a risk. Almost zero in this case - but don't sue me. LAVC is identical to "real" clusters, except for the communications medium (and HSCs). {What didn't they tell you - etc.} 1. (release notes has this - almost). When you boot (B) a satellite, it looks for SYSBOOT in every sysexe it can see locally. This includes all disks AND TKs and... if there's a tape in the tape drive, boot will read to EOT. This can take a long time - and cause boot timeout. Only after all of this does the satellite give up and try to get a load from Ethernet. Solutions are obvious, once you know. 2. Make sure you know the hardware ethernet address of the satellite. Various ways to find out - sys$hackers will just try to Boot the satellite from Ethernet. The Boot node will then print a message like "some node with hardware address xx-xx- just tried to get a load". VAX-31 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Then you have the info to execute satellite_config.com and add this node. 3. If all the satellites drop off the cluster, the Boot node tries VERY hard to find at least one satellite. Lots of Ethernet traffic, error messages, dropouts, etc. "Will be fixed in a future release". 4. VAXCLUSTER=2 is the keystone. To disable LAVC, just set VAXCLUSTER=0 (via Autogen, or Sysgen/write current) and reboot. 5. LAVC works really nicely. DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 707.6 Disk Serving for MicroVax? 6 of 6 "David Shepperd" 3 lines 21-AUG-1987 20:53 -< A MASS of LAVC disks >- ---------------------------------------------------------------- MASSBUS disks work just fine on a LAVC. We have both types in our LAVC and have had same since Jan '87 with no problems. So far as I can tell, floppy disks are the only disks not served by the MSCP. David Shepperd Atari Games Inc 675 Sycamore PO BOX 361110 Milpitas, Ca 95035-1110 (408) 434-1711 VAX-32 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 711.0 DMB32 Custom Device Driver Notes No replies "MARK SHAFFER" 27 lines 3-AUG-1987 14:18 ---------------------------------------------------------------- -< DMB32 Custom Device Driver Notes >- The following points amend Appendix B of the Release Notes for VMS V4.5, entitled "Generic VAXBI Device Support in VAX/VMS", for the case where the device is a DMB32: - When using SYSGEN CONNECT to load your driver, it is NOT sufficient to specify merely "/ADAPTER=". You must additionally specify "/NUMVEC=1 /VECTOR=128". Failure to do so will cause SYSGEN to crash the system. - In your device driver, specify "DMB32" rather than "GENBI" as the adapter type in the ADAPTER argument to the DPTAB macro. - My driver's controller initialization routine does a 5-second TIMEWAIT for the DMB32 to get itself together regardless of the state of UCB$W_POWER in UCB$L_STS. No nasty effects have been so far observed during loads by SYSGEN. MARK SHAFFER INFORMATION CONTROL TECHNOLOGIES,INC. 17 POLLY DRUMMOND PROF. BLDG. SUITE 105 NEWARK, DE 19711 302-366-8211 VAX-33 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 712.0 RIM 5 DATABASE No replies "JACK BERGER" 10 lines 3-AUG-1987 17:04 ---------------------------------------------------------------- HAS ANYONE DONE ANY MODIFICATION TO THE RIM 5 DATABASE PROGRAM CONCERNING THE REPORT GENERATOR? RIM 5 IS PUBLIC DOMAIN PACKAGE. AVAILABLE FROM NASA/COSMIC (NOW ON SIG TAPE). IN PARTICULAR I WOULD LIKE TO BE ABLE TO SPECIFY THE NUMERIC FORMAT FOR NUMERIC OUTPUT. AS SUPPLIED THE PROGRAM MAKES A BEST EFFORT ATTEMPT TO ROUND ANY NUMERIC DATA SO AS TO FIT IN A DEFAULT 7 CHARACTER (I BELIEVE) FIELD. IF I HAVE A NUMBER SAY .98765123, I WOULD LIKE TO BE ABLE TO OUTPUT AS .9877 FOR EXAMPLE. JACK BERGER FLUIDYNE ENGINEERING CORP. 5900 OLSON MEMORIAL HWY. MPLS. MN. 55422 612-544-2721 ================================================================ Note 715.0 New LUG ! No replies "MICHAEL GRATTAN" 11 lines 5-AUG-1987 10:14 ---------------------------------------------------------------- There is a new LUG for the Rhode Island, Southeastern Massachusetts and eastern Connecticut area. It's called the CRIMLUG and held it's second meeting on July 29. Anyone interested in more info can leave a note here or contact: Mr. Sam Whidden American Mathematical Society 201 Charles Street P.O. Box 6248 Providence, R.I. 02940 401-272-9500 MICHAEL GRATTAN FAIRHAVEN CORP. 358 BELLEVILLE AVE. NEW BEDFORD, MA. 02742 617-993-9981 EXT 106 VAX-34 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 716.0 ACC/ACL bug in RMS $OPEN 5 replies "Howard W. Pinsley" 41 lines 5-AUG-1987 19:21 ---------------------------------------------------------------- I ran into a problem with an application I had written in C that had run without incident for several months. Out of the blue I started getting and error return from an RMS $OPEN call of RMS$_ACC (I forget the full translation). The book said to check the secondary status which turned out to be something about a bad ACL format. This all seemed bizarre since all I was doing was an open (w/an XABDAT) to determine the file's dates -- I was doing nothing (or so I thought) with any ACL's. To make a short story long, I called Colorado support for help and lo and behold got it. Apparently there is a bizarre bug that can surface when an RMS operation (i.e. $OPEN) is invoked. It seems that the ACE chain must be traversed on every open to see if there are any alarm ACE's on the file. The rest I will quote from an explanation I received from Colorado. "The problem is the result of a path through the central protection checking routine where a register (R1) is not being correctly initialized. This happens as part of the search for security audit ACEs. The result is that when the check attempts to search the next ACL segment, it is using a bogus value in R1. The results will vary depending upon what happens to be in memory after the ACL segment. The only way to see this problem is to have a file that requires two or more ACL segments. Since each segment comes from a file header, the file must be multi-header. Thus if you have a multi-header file, with ACL information in more than one segment, you may see this problem. This will be fixed in V4.7" Has anyone run into this problem other than me. The guy I was speaking with added that we could see the problem rear its ugly head virtually anywhere since it was problem with RMS. Despite the fact that it seems to be a relatively rare error, it scares me. Also, it would seem a pervasive problem like this warrants a patch - after all, 4.7 is still pretty far away. VAX-35 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Any thoughts? Howard W. Pinsley 145 West 79th. Apt 11D New York, NY 10024 ================================================================ Note 716.1 ACC/ACL bug in RMS $OPEN 1 of 5 "Lee K. Gleason" 11 lines 9-AUG-1987 03:30 -< Hurts Me Too >- ---------------------------------------------------------------- Happens at one of my sites all the time - environment there is COBOL. We realized it was a multi header file with ACL info in more than one of them problem, and deal with it by creating a new, single header file. We didn't call the problem in to Colorado or write an SPR, since we guessed that we would get the helpful answer you did. Lee K. Gleason Control-G Consultants 2416 Branard #D Houston TX 77098 713/528-1859 ================================================================ Note 716.2 ACC/ACL bug in RMS $OPEN 2 of 5 "Howard W. Pinsley" 2 lines 11-AUG-1987 12:51 -< Multi/Single Header Files >- ---------------------------------------------------------------- I guess that solution will have to do for the time being. How do make a multi-header file single-header? Howard W. Pinsley 145 West 79th. Apt 11D New York, NY 10024 VAX-36 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 716.3 ACC/ACL bug in RMS $OPEN 3 of 5 "Larry Kilgallen" 11 lines 11-AUG-1987 19:42 -< Getting rid of multiple headers on files >- ---------------------------------------------------------------- If the cause of multiple headers is fragmentation (too many extents due to fragmented free space or gradual appending), it will be cured when you do your regular image backup and restore to defragment the disk. If the cause of multiple headers is excessive ACL entries you will have to produce a site-specific solution. (I maintain that systematic ACL application at a site should always be based on general identifiers (e.g., THEOCRACY_USER) rather than specific UICs. In most cases that will make the list shorter. Larry Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 ================================================================ Note 716.5 ACC/ACL bug in RMS $OPEN 5 of 5 "Howard W. Pinsley" 5 lines 16-AUG-1987 18:17 -< Its looking worse >- ---------------------------------------------------------------- Some more info on the original problem: The bug is now surfacing in standard VMS functions such as copy and backup. Notably our nightly backups are starting to get errors and files are not getting backed up. Howard W. Pinsley 145 West 79th. Apt 11D New York, NY 10024 VAX-37 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 717.0 CR to Fortran Carriage Control 8 replies "ROGER G. RUCKERT" 9 lines 6-AUG-1987 10:09 ---------------------------------------------------------------- We have a VAX 8500 and want to convert a file with carriage return carriage control to Fortran carriage control. "Convert" and "FDL" say they should be able to do it, but they only go the other way. Colorado Springs doesn't know what else to try. Someone in our DECUS LUG suggested using TECO, but I understand that product doesn't run on 8xxx hardware (no compatability mode). Has someone written a conversion program somewhere (e.g., DECUS tapes) or do I have to write my own? ROGER G. RUCKERT MEDTRONIC, INC./MAIL STOP B100 7000 CENTRAL AVE. MINNEAPOLIS, MN 55432 (612)574-4742 ================================================================ Note 717.1 CR to Fortran Carriage Control 1 of 8 "Brian Tillman, Lear Siegler, Inc." 8 lines 6-AUG-1987 22:09 -< Aw, you could do it yourself! >- ---------------------------------------------------------------- RE: < Note 717.0 by NODE::US217557 "ROGER G. RUCKERT" > -< CR to Fortran Carriage Control >- Has someone written a conversion program somewhere (e.g., DECUS tapes) or do I have to write my own? Writing your own would be fairly trivial. I'd estimate about 50 lines tops. Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 VAX-38 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 717.2 CR to Fortran Carriage Control 2 of 8 "John Burnet" 27 lines 7-AUG-1987 12:24 -< Here's one solution >- ---------------------------------------------------------------- No need to write a program... try this. (Assume that you have an existing RAT=CR file, FILE.ONE, and you want to create a new RAT=FTN file, FILE.TWO.) ----------------------------------------------------- $ create/fdl=sys$input file.two file organization sequential record carriage_control fortran format variable $ copy/overlay file.one file.two;0 ----------------------------------------------------- Ignore the "%COPY-W-INCOMPAT" incompatible-attributes message. However, the first character of each line in the existing RAT=CR file *WILL* be used as a carriage-control specifier in the newly created RAT=FTN file. If you want to convert a RAT=CR file to a RAT=FTN file and actually *add* the carriage-control bytes, you'll need to cons up a little program to do it. I'd say that 50 lines is a gross over-estimate, unless you tend to use a lot of gotos branching around gotos... :-) I estimate that about 15 lines of Fortran will do it, unless you want to get fancy and translate "newline, record" and "newline, newline, record" sequences into "0record" and "-record" -- that is, add the "0" and "-" carriage control characters while throwing away the null lines they replace. John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 VAX-39 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 717.4 CR to Fortran Carriage Control 4 of 8 "Frank J. Nagy" 10 lines 8-AUG-1987 10:32 -< See REFORMAT from VAX SIG tapes >- ---------------------------------------------------------------- I once found a utility on the VAX SIG tapes called REFORMAT which does just what is wanted here - along with lots of other conversions of file formats. REFORMAT will also fix up the records in converting from CR to FORTRAN record format. Sorry, I don't remember which tape REFORMAT is on, but since I've had it around awhile it probably first appeared no later than Spring '85 at a guess. Try looking on the 83/84 tapes for a start. REFORMAT is another of those useful utilities you wish Digital would provide... Frank J. Nagy Fermilab PO Box 500 MS/220 Batavia, IL 60510 (312)840-4935 ================================================================ Note 717.5 CR to Fortran Carriage Control 5 of 8 "Kevin Angley" 4 lines 9-AUG-1987 17:35 -< Where to find REFORMAT >- ---------------------------------------------------------------- I believe you will find REFORMAT packaged with VAXNET on the VAX86A tape. Kevin Angley 3301 Terminal Drive Raleigh, NC 27604 (919) 890-1416 VAX-40 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 717.6 CR to Fortran Carriage Control 6 of 8 "Brian Tillman, Lear Siegler, Inc." 18 lines 10-AUG-1987 14:12 -< I can write that code in xx lines. >- ---------------------------------------------------------------- RE: < Note 717.2 by NODE::US124053 "John Burnet" > -< Here's one solution >- | I'd say that 50 lines is a gross over-estimate, unless you tend | to use a lot of gotos branching around gotos... :-) Absolutely not a gross overestimate. You haven't allowed for the overhead of obtaining the parameters (one input, one output) with the CLI interface routines, the OPEN and CLOSE statements, the calls to ERRSNS and LIB$SIGNAL in the event of an I/O error and a few other factors needed to make the program bullet-proof. Sure, quick-and-dirty would only need 15 lines, but the days of quick-and-dirty are past and if that's the way a person writes his/her code, some education is in order. The program I had in mind, even with the proper structure, would still only need 20 minutes of coding and wouldn't have to change again in it's life time, whereas one "cons"ed up for a specific case would have to be tossed out with the garbage as soon as the file name changed. Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 ================================================================ Note 717.7 CR to Fortran Carriage Control 7 of 8 "Bruce Bowler" 6 lines 10-AUG-1987 16:55 -< Nothing wrong with occasional hacks >- ---------------------------------------------------------------- The original note stated that they had A file (a implies 1) to convert. In that situation there is absolutely no need to write 'bullet proof' code. While it is true that people who write shoddy production code should be hung by their (fill in your favorite body part), there is nothing wrong with a good ol' hack every now and then. Some of my best code started out that way. VAX-41 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Bruce Bowler General Electric 1 River Road Bldg 2 Room 609 Schenectady, NY 12345 ================================================================ Note 717.8 CR to Fortran Carriage Control 8 of 8 "John Burnet" 49 lines 19-AUG-1987 12:46 -< n lines of code vs. n * 5 lines >- ---------------------------------------------------------------- Re. .6: -- > Sure, quick-and-dirty would only need 15 lines, but the days of > quick-and-dirty are past and if that's the way a person writes > his/her code, some education is in order. My, my, aren't we touchy. [Programmers, that is... ;-)] Lest it be thought that I'm in the habit of cranking out hackified code that breaks when the moon enters its next phase, let me clarify. First, I agree with Bruce (see .7) that people who write shoddy production code should be banished to the wilderness, never to dial 1-617-262-6830 again. Actually, I think that the banishment should extend to people who write shoddy *any kind of* code, but the wilderness would become a little crowded... Second, most throwaway programs don't need to be bullet-proof, and making them so is usually a waste of time (unless you learn some- thing useful in the process). It *is* important to know what programs you write are actually going to be "one-use" code and which ones are--to paraphrase Kernighan and Plauger--likely to be pressed, bent, hammered and twisted into years of service. I think that most of us try to fool ourselves somewhat in that regard. ("Aw, I'm the only one who's ever going to use the thing; who cares what it looks like?") I, too, like to trap all possible errors, handle command-line arguments, etc. in my utility programs. However, if you have a stable full of utilities whose source contains a lot of the same code for doing this kind of thing, I suspect that you need to factor out the common operations into subprograms that you can call as needed. For example, instead of repeatedly inserting VAX-42 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT calls to errsns and lib$signal, it would probably be better to have something like: read (f, '(q, a)', iostat=status) length, line if (status .ne. 0) call handle_fortran_error (status) or even just "call lib$establish (handle_fortran_error)" (with the handler's argument list declared appropriately). After a while you build up a library of these routines, and using them becomes second nature. *Then* you can write 15- and 20-line main program modules (for something as trivial as a read-record/ write-record loop, that is) and still have your error checking too. Maybe we should move this discussion to a new topic: programming style flames. John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 ================================================================ Note 718.0 Counting Users 5 replies "ROGER G. RUCKERT" 9 lines 6-AUG-1987 10:15 ---------------------------------------------------------------- We want to be able to tally the number of interactive users on the system at any one time. Rather than having a job sample every so many minutes, I thought I could just read the accounting log files and calculate from users' signon and signoff times how many people were signed on. Does anyone know of an easier way to do this from the accounting files? Is there another approach to this whole problem? Finally, has anyone already written a program that will actually do the tallying? Thanks much. ROGER G. RUCKERT MEDTRONIC, INC./MAIL STOP B100 7000 CENTRAL AVE. MINNEAPOLIS, MN 55432 (612)574-4742 VAX-43 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 718.1 Counting Users 1 of 5 "Brian Tillman, Lear Siegler, Inc." 14 lines 6-AUG-1987 22:17 ---------------------------------------------------------------- RE: < Note 718.0 by NODE::US217557 "ROGER G. RUCKERT" > -< Counting Users >- | We want to be able to tally the number of interactive users on | the system at any one time. Aw, heck, wouldn't be easier to just say $ SET LOGINS every now and then and record the answer? I know it's human intensive, but you could do this many, many times (at about 1 second per time) before you'd use up the time it would take you to write a program to handle the accounting file. I mean, automation is great, but a pencil and paper are still useful tools when you know the appropriate time to use them. Never use a pile driver when a tack hammer will do. Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 ================================================================ Note 718.3 Counting Users 3 of 5 "Dale E. Coy (505) 667-3270" 37 lines 23-AUG-1987 00:49 -< How about this? >- ---------------------------------------------------------------- How to find out the number of interactive and batch users: Extracted from program WHERE, by Andrew W. Potter (AT&T, I think), Nov 25th, 1985 - someWHERE on the SIG tape. FORTRAN (of course) Subroutine Interactive (Interactive, Batch) Implicit Integer*2 (a-z) Integer*4 Interactive, Batch External Sys$GW_Ijobcnt External Sys$GW_Bjobcnt VAX-44 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Int = Peek (Sys$GW_Ijobcnt) Bat = Peek (Sys$GW_Bjobcnt) Return End Function Peek (I) Integer*2 I, Peek Peek = I Return End We have this in a larger program which tells the user a lot more stuff, such as node name, version of VMS, terminal name, etc. The program is compiled and linked with the following: $ Fortran/nodebug/nocheck Programname $ Link/nodebug/notrace Programname,Sys$System:sys.stb/sel I didn't take the time to figure out whether all of this was necessary - but trust you can take it from here. It's really a very nice hack - even if the code is a bit obscure. Let's hear it for the SIG tapes, and Pageswapper. DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 718.5 Counting Users 5 of 5 "Ron Schneider 312-498-5615" 8 lines 26-AUG-1987 00:43 -< Interactive job count from DCL >- ---------------------------------------------------------------- If you are lazy like I am, you can do the same thing from DCL. $ INTERACTIVEJOBS = F$CVSI(0,16,F$FAO("!AD",4,%X80002BFC)) $ BATCHJOBS = F$CVSI(0,16,F$FAO("!AD",4,%X80002300)) VAX-45 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT with all of the normal provisos about the address changing with a new release. You can actually access all of the URKW data detailed in SYS.MAP with this method. ronald e. schneider odesta corp, world headquarters 4084 commercial drive northbrook, il 60062 312-498-5615 ================================================================ Note 719.0 BACKUP status codes 2 replies "John McGrath" 10 lines 7-AUG-1987 01:26 ---------------------------------------------------------------- Why does BACKUP (VMS V4.5) return status codes that are not in the system message file? Does anybody have a list of the associated messages? Some of the status codes I have gotten are %X10A38OFC, %X10A38OF4, and %X10A3815C. John P. McGrath Software Consulting Services 3162 Bath Pike Nazareth, PA 18064 (215) 837-8484 ================================================================ Note 719.1 BACKUP status codes 1 of 2 "Saul Tannenbaum" 16 lines 7-AUG-1987 17:30 -< Backup status codes >- ---------------------------------------------------------------- > Why does BACKUP (VMS V4.5) return status codes that > are not in the system message file? When I asked that question of CSC a long time ago, the answer was that backup (when used as stand-alone backup) needed to be able to run in environments when there was no system message file. That's more than reasonable, but doesn't answer the question as to why they aren't _Also_ in the system message file for convenience. Saul Tannenbaum Tufts University HNRC VAX-46 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT 711 Washington Str. Boston, MA 02111 (617)556-3346 ================================================================ Note 719.2 BACKUP status codes 2 of 2 "Larry Kilgallen" 5 lines 8-AUG-1987 09:09 -< Try SET MESSAGE SYS$MESSAGE:SYSMGTMSG >- ---------------------------------------------------------------- I think message handling in this regard leaves a LOT to be desired; one should not have to issue an extra command just to find out what an error message is in DCL (which is where a lot of us do it to send mail with f$message, etc.). Larry Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 ================================================================ Note 720.0 TPU CALLUSER procedures 6 replies "John McGrath" 27 lines 7-AUG-1987 01:27 ---------------------------------------------------------------- Has anybody had any luck with TPU "calluser" procedures? 1) I have tried to write one in VAX Pascal, but it does not allow me to specify string descriptor parameters for anything but external procedures. 2) I have tried writing a VAX C version. It works fine, except that the string descriptor for ret_string gets messed up after calluser returns. The data type field of the descriptor is changed to 0. TPU blows off when it tries to change change the value of the returned variable. 3) Is there any way that I can write a procedure that can be called by TPU invoked by MAIL without placing it in SYS$LIBRARY and installing it? I do not mind the install, but I would rather that it not go in SYS$LIBRARY. VAX-47 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT I'd like to here from anybody who has used CALLUSER. John P. McGrath Software Consulting Services 3162 Bath Pike Nazareth, PA 18064 (215) 837-8484 ================================================================ Note 720.1 TPU CALLUSER procedures 1 of 6 "Mark Oakley" 17 lines 9-AUG-1987 22:16 -< "ret-string" is dynamic >- ---------------------------------------------------------------- TPU passes "ret_string" via a dynamic-string descriptor. I think this is documented somewhere in the TPU RM, but I'm not sure where. I wrote my TPUCALLUSER image in Fortran, and use the RTL routine "STR$COPY_DX" to return a character string. (The example in the TPU RM is coded in VAX BASIC, which automatically generates calls to the appropriate RTL routines to handle static and dynamic character strings.) I have not had any success in getting TPUCALLUSER to be invoked while I'm in MAIL. What sorts of things are you doing in your TPUCALLUSER? I use mine to show/set my default directory and "pass- thru" control-keys to my TPU procedures. (They should be on the VAX SIG Nashville tape.) Mark Oakley Battelle Memorial Institute 505 King Ave. Columbus, Ohio 43201-2693 614/424-7154 VAX-48 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 720.2 TPU CALLUSER procedures 2 of 6 "Bruce Bowler" 10 lines 10-AUG-1987 16:48 -< Use logical names >- ---------------------------------------------------------------- 3) Is there any way that I can write a procedure that can be called by TPU invoked by MAIL without placing it in SYS$LIBRARY and installing it? I do not mind the install, but I would rather that it not go in SYS$LIBRARY. Try setting up a logical name to point to where the sharable image is for your procedure. for example DEFINE TPUCALLUSER dev:[dir]TPUCALLUSER Bruce Bowler General Electric 1 River Road Bldg 2 Room 609 Schenectady, NY 12345 ================================================================ Note 720.3 TPU CALLUSER procedures 3 of 6 "John McGrath" 14 lines 10-AUG-1987 22:51 ---------------------------------------------------------------- I have had no problem with CALLUSER from MAIL; The only thing is you have to install it because MAIL is a privileged image. So when you are running your TPU procedure, you are privileged. I had intended to use CALLUSER for a number of things. First is for some set directory/show directory commands. Next I'd like my TPU procedure to know where it is being run from, i.e. MAIL, etc. I've been doing this based on the file name associated with the MAIN buffer on startup, but I rather do it from the image name. I'd also like to use TPU for some other applications that are primarily text oriented, but that require some fancier file access than TPU can handle on its own. John P. McGrath Software Consulting Services 3162 Bath Pike VAX-49 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Nazareth, PA 18064 (215) 837-8484 ================================================================ Note 720.4 TPU CALLUSER procedures 4 of 6 "John McGrath" 16 lines 11-AUG-1987 00:11 -< Doesn't work for me >- ---------------------------------------------------------------- Try setting up a logical name to point to where the sharable image is for your procedure. for example DEFINE TPUCALLUSER dev:[dir]TPUCALLUSER I have tried that, but I get: Error activating image DUB1:[SYS0.][SYSLIB]TPU$CALLUSER.EXE; File not found Could not find a routine for CALL_USER to invoke I have also tried defining TPU$CALLUSER, but no luck. Does the logical name assignment have to be in executive mode or have any special /TRANS= ? John P. McGrath Software Consulting Services 3162 Bath Pike Nazareth, PA 18064 (215) 837-8484 ================================================================ Note 720.5 TPU CALLUSER procedures 5 of 6 "Chris Erskine" 7 lines 14-AUG-1987 08:05 -< To use CALL_USER >- ---------------------------------------------------------------- The install of the TPU$CALLUSER file requires the logical TPU$CALLUSER to be defined. This is documented in the Guide to Creating Modular Procedures on VAX/VMS. I am using CALL_USER to create my own LSE like editor remembers some things in logical names such as name of file and where in file I was last editing. Chris Erskine 6001 Adams Rd. VAX-50 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Bloomfield Hills, MI 48013 (313) 258-4049 ================================================================ Note 720.6 TPU CALLUSER procedures 6 of 6 "John Burnet" 11 lines 19-AUG-1987 12:18 -< TPU$CALLUSER logical name definitions >- ---------------------------------------------------------------- Re: .4, failure of TPU to find TPU$CALLUSER logname translation In certain circumstances, RMS will only use what it considers to be "trusted" logical names, which are those defined in the system table in exec mode. Those "certain circumstances" include the case of an image installed /protected, and may also--I'm not sure about this--include images installed with privileges and/or installed images with only execute permission set on the file. Just for kicks, try using "DEFINE/SYS/EXEC TPU$CALLUSER whatever" and see if that solves the problem. John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 ================================================================ Note 721.0 SET UIC from an image 4 replies "John McGrath" 26 lines 7-AUG-1987 01:28 ---------------------------------------------------------------- Does anybody have any idea how to do the equivalent of a SET UIC from within an image? I would also like the UIC to revert the original on image exit (sort of a user mode SET UIC). I think this would be the best way to handle security for our application, but maybe I'm taking the wrong approach. Our application has it's own security checks built in to it. The criteria for determining whether access to a given object should be granted is complex enough that VMS's UIC based protection will not work, and ACLs probably wouldn't either. At best they would be overly complex. VAX-51 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT So what I would like to do is to have all of the files, queues, etc., belong to a UIC assigned to the application. The application programs would change to the application's UIC when it started. It would determine whether to grant access to any object. When it is done, it would revert to the user's UIC and return control to the user. John P. McGrath Software Consulting Services 3162 Bath Pike Nazareth, PA 18064 (215) 837-8484 ================================================================ Note 721.1 SET UIC from an image 1 of 4 "John Osudar" 11 lines 7-AUG-1987 21:29 -< easy >- ---------------------------------------------------------------- You can do anything you want from kernel mode... I have a subroutine sitting around somewhere that sets the UIC. You can't very well make it revert to the original on image exit, unless you have the image trap its exit with an exit handler, and call the routine again at that point. Otherwise, it's simple to do, and applicable to other process/job parameters; e.g. I have a routine that sets the username, to be used only in very, VERY privileged code!!! :-) If you want the code (it's actually easy to figure out how to do it yourself, using SYS$CMKRNL and a copy of the process/job data structure definitions) I can mail it to you on this system... John Osudar Argonne National Laboratory 9700 S. Cass Ave. Bldg. 205 A-051 Argonne, IL 60439-4837 (312) 972-7505 VAX-52 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 721.3 SET UIC from an image 3 of 4 "Mark Oakley" 15 lines 9-AUG-1987 22:23 -< User System Service for Changing UIC >- ---------------------------------------------------------------- I have written a pair of system services that temporarily change the process uic to that of the image-file uic. One service changes the uic, and the other restores it. A run- down handler is included, so that the process uic is restored if the restore service is not executes. Control-Y and control-c handlers are included to make sure that a user can not escape to DCL with the wrong uic. Mark Oakley Battelle Memorial Institute 505 King Ave. Columbus, Ohio 43201-2693 614/424-7154 ================================================================ Note 721.4 SET UIC from an image 4 of 4 "John Burnet" 31 lines 10-AUG-1987 12:50 -< Why not just remove w access & install w/sysprv? >- ---------------------------------------------------------------- The set-uic routines are definitely possible (even easy if you know what you're about), but why bother? Here's my two cents... First, there's probably no point in trying to protect files and queues from anyone with a system UIC or SYSPRV or BYPASS privilege, since those users can simply change the protection to allow any access they want. (Obviously, this also applies to several other privileges like CMKRNL, SYSNAM, SETPRV, etc.) So all of your users with (actual or effective) system UICs should be trusted not to mess up whatever scheme you decide to put into place. With that limitation, it seems to me that the easiest way to do what you want is simply to have the files and queues owned by a system UIC, remove *all* world access from those objects, and install with SYSPRV privilege all programs which use those files and queues. (BTW, remember that you can effectively "install" a command procedure with privileges by disallowing world access to the procedure and installing with SYSPRV an executable image VAX-53 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT which simply uses lib$spawn to run the procedure in a subprocess. Users invoke the image, which can be installed with privileges other than SYSPRV and which will pass all such privileges on to the subprocess.) The "install with SYSPRV" scheme is, as far as I'm aware, the closest that VMS comes to Unix's "setuid" bit. Of course, you could also create your own "setuic" system service, code it to require CMKRNL, and install all of your programs for this application with CMKRNL. I myself don't see the point of doing that when all you really need to protect against is access by non-system uics. (Unless you want to keep close track of resource accounting for the application, using the UIC to track usage; but you can always enable image accounting for installed images anyway.) John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 ================================================================ Note 722.0 Long Distance Data Communications No replies "MICHAEL GRATTAN" 16 lines 7-AUG-1987 07:29 ---------------------------------------------------------------- We are looking at various ways of sending data to out west coast warehouse, we are on the east coast. Our applications call for them to be on-line as opposed to batch transmissions. We are searching for alternatives to leased lines and are thinking about going via satellite. (Very Small Aperture Terminals, to be specific.) The cost are about the same, as far as we can tell. Does anyone have any experience using satellite communications for data? If so, how reliable and clean are the 'lines'? Does anyone have any other suggestions besides satellites, leased lines or carrier pigeons? Any thoughts would be greatly appreciated. Thanks. MICHAEL GRATTAN FAIRHAVEN CORP. 358 BELLEVILLE AVE. NEW BEDFORD, MA. 02742 617-993-9981 EXT 106 VAX-54 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 723.0 VMS to a Mini-Exchange 1 reply "Keith Roberts" 15 lines 7-AUG-1987 10:11 ---------------------------------------------------------------- Our department has purchased a DEC Mini-Exchange Box. This was bought to allow a DECmate III and our Vax to share the DECmate's LN03. DECmate WPS seems to support the Mini-Exchange very well...However it appears that the VMS print symbiont will require some changes; to establish communications to the Box, and to properly disconnect...And maybe even some error checking. Does anyone out there know of any modules already developed to allow VMS to join in with a Mini-Exchange? Thanks, Keith Roberts SASC Technologies Inc 109 Mass Avenue Lexington, Ma 02173 (617) 377-5959 ================================================================ Note 723.1 VMS to a Mini-Exchange 1 of 1 "Bill Mayhew" 28 lines 10-AUG-1987 12:25 -< Another hardware solution? >- ---------------------------------------------------------------- I don't know of such a thing, but if you find it elsewhere, please let me know here. I've written MiniExchange drivers for other operating systems on 11's and VAXen. Basically the issue is that the VAX needs to toggle Data Terminal Ready in order to gain the MX's attention, then send it a short stream of data to select the appropriate port (to which the printer is connected), then monitor the serial line for a response from the MX as to how successful it is. I know nothing about modifying the print symbiont so I can't even guess as to whether that can all be done. (Disconnecting just means dropping DTR, and there is another control signal that can be watched (DSR? I forget) that will tell you if it succeeded.) VAX-55 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT There is a hardware fix, however. A company called Gold Key Electronics, somewhere in New Hampshire, makes a box that does what you need simply by watching the incoming lines for *data* (no control signals required). It is, on the other hand, strictly a printer sharer. My use for the MX involved a single serial line on a CPU that might be used for outgoing print jobs to one of several printers, *or* for incoming dialups from a modem attached to the MX. Getting VMS to support that functionality would be more of an exercise :-). Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 ================================================================ Note 724.0 Revision F1 of DMZ32 fixes several problems 1 reply "Jonathan Pinkley" 137 lines 7-AUG-1987 19:51 ---------------------------------------------------------------- After we received our VAX 8600 in November of 1985, we noticed that the interrupt from a control Y was not handled by the DMZ32 in the same manner as it was on our DMF32 or our (ABLE) VMZ32. In fact at low baud rates, characters from other users processes sometimes appeared after the ^Y was pressed. This problem only occurred on the DMZ32. We observed that this behavior was prevented by turning off DMA mode. We set the default to /NODMA for our terminals via the sysgen parameter TTY_DEFCHAR2. We contacted Colorado Springs and they were "not able to reproduce" the problem. They did observe the problem when we asked them to dial into our system while we had our line monitor on the modem. This was reported early in 1986. Soon after I had reported this to CSC, I talked with Gary Grebus, who was then working at Battelle. Battelle had a configuration similar to ours. Gary was able to reproduce the problem on Battelle's DMZ32's. After we received VMS V4.4 we again noticed the problem. We traced this to the fact that VMS V4.4 ignores the setting of the DMA bit in TTY_DEFCHAR2. We spr'd this and wrote a program to set the terminal mode during our startup. VAX-56 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT On July 27, 1987 I again called the customer support center. This time I reported it as a security problem (since I could easily get data that was not meant for me to appear on my screen). I gave them the procedure I used to cause the problem to appear and this time they were able to reproduce the problem. CSC searched their private database and found that there was a recent FCO for the DMZ32. I got the FCO number so that I could give it to our field service engineer. We had our DMZ32's upgraded on August 4, 1987 and this does fix the problem. I have to complement DEC on the FCO, it fixes problems that people did not even find. What bothers me is that: 1. The Colorado Customer Support Center terminal support team was not aware of the FCO and what it fixed. 2. The information was in their database, but is still not in the DSIN customer database (as of 8/6/87) (there is a note on how disable the self test of the DMZ DMA and SPEED so that the device will autoconfigure. I haven't tried changing these since the FCO was installed.) 3. Field service was not aware of the availability of the FCO and our Field service engineer had to give the FCO kit numbers to district before they were able to track it down. Lessons learned: 1. Digital seems to pay more attention to calls that are security related. 2. If you get any information from CSC make sure you get FCO order numbers so that your field service engineer will be able to get the parts. 3. Don't count on DEC to spread the word about fixes. 4. Don't expect to find any timely information on DSIN. (At least now you can access DSIN via TYMNET so you won't have to pay extra for the phone call.) VAX-57 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT So you won't have to go through the bother I did, below are some excerpts from the paper that came along with the new boards. Order by FCO Qty Part numbe Description Kit # EQ-01457-01: 1 M8398 Module at Revision F1 FA-04743-01: 1 FCO Document Extracted from FA-04743-01. (FCO DMZ32-I001 documentation) Applicability: Replace Revision E1 or earlier M8398 module with a Revision F1 or later to incorporate new UBI microcode version 16 and TIU microcode version 1C when problem/symptoms are evident. This FCO implements ECO M8398-MK003. Problem/Symptoms: 1. Problem: Flush bit not clearing DMA correctly Symptom: File corruption on unibus disk controller drives and some users getting other users data. System error logger logs "IMR" invalid map register errors. This bug requires that "DMA" mode be turned off on all DMZ32 lines on the system. This is a VAX/VMS specific problem. 2. Problem: SSYNC timeout counter starts before NPG is received [ NXM error ]. Symptom: This causes Ultrix and UNIX systems to hang and or crash based on operating system environment at the time of the timeout. 3. Problem: When H3014 is powered off/on, DTR is not reasserted by microcode. Symptom: Terminal lines set up as modem will not be able to connect to system, as well as users logged in through modem connections will be dropped due to no "DTR on" status. System must be rebooted to re-enable modem lines. 4. Problem: Unsupported baud rates cause microcode to hang. VAX-58 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Symptom: If un-suspecting user tries to set terminal to one of the two unsupported baud rates, the DMZ32 microcode will hang and the device will be unusable to the system. A total system reboot is required to clear the condition. 5. Problem: NXM error sets the wrong bit in Octet CSR. Symptom: No error reports from field, but if NXM were to occur, the YCDRIVER would not be notified of it's occurrence. This could have serious implications on 8XXX system with DWBUA [unibus adapter]. 6. Problem: Self test code constants enabled at wrong time. Symptom: None reported from field, but was discovered during internal testing of DMZ32 microcode. 7. Problem: Foreign terminals sending abutted characters lose trailing bits. Symptom: Files get invalid escape sequence data while editing using the "arrow" cursor positioning keys. Jonathan Pinkley Gould OSD Dept. 913, Bld. 2 18901 Euclid Avenue Cleveland, OH 44117 (216)486-8300 x1335 ================================================================ Note 724.1 Revision F1 of DMZ32 fixes several problems 1 of 1 "Wayne L Dernoncourt" 7 lines 29-AUG-1987 19:34 -< FCO fix >- ---------------------------------------------------------------- See the notes in 663 [ Pageswapper Volume 9 Number 1 ] for more information and complaints about DEC not always knowing what's going on. I first heard this problem mentioned in a LUG meeting in January(?) 87 meeting, the gentleman said that yes DEC knew about the problem, but didn't intend to fix it because it was a mature product!! I heard about the FCO in Nashville at the Advanced Q & A Session (actually the session before that one). VAX-59 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Wayne L Dernoncourt NAVAL ORDNANCE STATION C/O 1209 Bannister Circle WALDORF MD 20601 ================================================================ Note 725.0 LN03-Plus vs. LN03R ScriptPrinter 15 replies "Brian Tillman, Lear Siegler, Inc." 9 lines 10-AUG-1987 15:04 ---------------------------------------------------------------- We are purchasing a VT330 for a MicroVax for the purpose of creating ReGis graphics. After the pictures are correct, we then want to be able to print the graphics on an LN03. Perhaps someone can tell me: Is the LN03-plus enough or do I need the ScriptPrinter (LN03R) and it's ReGis to Postscript filter? My sales representative says I need the LN03R (and when I read the VAX Systems and Options Catalogue, it appears he may be correct), but I've heard rumors the LN03-Plus will do the job (and save me a couple grand). Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 ================================================================ Note 725.1 LN03-Plus vs. LN03R ScriptPrinter 1 of 15 "Bill Mayhew" 26 lines 10-AUG-1987 17:16 -< It depends... >- ---------------------------------------------------------------- I think the question comes down to how you want to print the graphics. If you want to print them directly from ReGIS commands generated on the (micro)VAX, then your representative is probably correct (save for some possible DECUS ReGIS-to-Sixel software I keep hearing rumors about). If, on the other hand, you either (a) attach the LN03-Plus to the VT330's printer port and print the graphics through the VT, or (b) have the VT upload the graphics back to the host, in sixel form, to a file which can then be printed on the LN03-Plus, you should be home free. VAX-60 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT DECslide and DECgraph, btw, work by using this reverse-direction hack. It seems the VT3xx (and VT24x) can convert what's on their screen into sixels and send it either out the printer port, or out the comm port back to the host. The base LN03 and the -Plus both understand sixels, but neither understands native ReGIS. Btw, there is an upgrade for the 03 (or 03-Plus) to the 03R, which I understand is "bargain priced" through the end of the year, when it will go up. Last time I looked it was cheaper to buy an 03, plus the upgrade kit, than an 03R. Finally, I have seen in the marketing literature some obscure references to a possibility that some of the functionality of the 03 and/or 03-Plus "goes away" when you go to the 03R. Note, for instance, that the new "script-style" font cartridge listed in the latest DECdirect-Plus specifically carries a warning that it won't work with the 03R. Does anybody know exactly what the tradeoffs are between these beasts? Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 ================================================================ Note 725.2 LN03-Plus vs. LN03R ScriptPrinter 2 of 15 "Jack Patteeuw" 38 lines 10-AUG-1987 18:37 -< More input >- ---------------------------------------------------------------- Please note that this is all "second hand" information ! The LN03R is a Postscript device. That means it only speaks Postscript ! It does **NOT** understand ANY of the sequences that you normally use to do neat things on either the LN03 or the 03+. Consequently, any output that is not in Postscript form must be run through some kind of a filter program to convert it to Postscript (this includes ReGIS, Sixel, and plain old LN03 command files). If your graph is complex this could take some time and CPU ! VAX-61 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT re .1 Yes you can hook either a LN03 or a 03+ to the printer port of the VT24x or VT3xx and get Sixel pictures, but be aware that these are of very LOW QUALITY (use a LA75 or some thing else cheap for Sixel output). The software to convert ReGIS to Sixel was published in the Wombat Examiner within the past few months. It does require you to have a spare VT24x or VT3xx to do the conversion. I think it is the DECSlide product that converts ReGIS to Sixel. Finally, another option would be to buy the LN03+, which does understand TEK 4010/14 (and works quite well from my short testing). Then, I think you can use a product called Baseview to convert ReGIS file to TEK 4014 (important: you want 4014, not 4010. 4014 has 4 time the resolution !) and then print it on the LN03+. Anyway you do it, put it in writing with DEC that it will work and will be supported. If not cancel the deal. I don't know how much pull your company has with DEC but you should be able to get the to DEMONSTRATE the solution at your site or theirs (with your data of course) before you give them an order. Jack Patteeuw Ford Motor Co. Electrical and Electronics Division 31630 Wyoming Livonia, MI 48150 313-323-8643 ================================================================ Note 725.3 LN03-Plus vs. LN03R ScriptPrinter 3 of 15 "Larry Kilgallen" 1 line 10-AUG-1987 20:12 -< NO font cartridges work with the LN03R >- ---------------------------------------------------------------- It is strictly a PostScript machine (in response to 725.1). Larry Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 VAX-62 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 725.4 LN03-Plus vs. LN03R ScriptPrinter 4 of 15 "Brian Tillman, Lear Siegler, Inc." 8 lines 11-AUG-1987 10:34 -< Some clarification >- ---------------------------------------------------------------- What we intend to do is generate a ReGis file on disk and copy it to the VT330, modifying it until it's correct (or just generate it on the VT330 until it looks correct and then generate a file). This file, then will be passed through the LN03R ReGis to Postscript filter program (which we also will order). The printer must be used for regular output as well, with a queue pointing to it, so chaining it off the VT330 is probably not what we wish to do. Brian Tillman Lear Siegler, Inc. 4141 Eastern Ave. MS121 Grand Rapids, MI 49518-8727 (616)241-8425 ================================================================ Note 725.5 LN03-Plus vs. LN03R ScriptPrinter 5 of 15 "Bob Hassinger" 17 lines 12-AUG-1987 10:02 -< Conversion, not upgrade >- ---------------------------------------------------------------- At the Spring Symposium it was made very clear to me by DEC people that the LN03/LN03+ to LN03R kit is NOT an "upgrade" kit. It is a "conversion" kit. By that, DEC means you are not *adding* capabilities, you are *changing* from one set to another. In the case of VMS with the translator software package you end up with a similar overall capability at some (unknown) processing cost but the LN03R will not be useful at all connected to say a DECmate the way the others are because DECmate WPS speaks only ASCII and DEC escape sequences, not Postscript. I found this disappointing. I wonder how much more it would have taken to include an LN03 compatability feature for ASCII files. I have heard some other Postscript printers in this market do have a mode for ordinary ASCII. Bob Hassinger Liberty Mutual Research Center VAX-63 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 725.6 LN03-Plus vs. LN03R ScriptPrinter 6 of 15 "Frank J. Nagy" 8 lines 15-AUG-1987 11:17 -< ASCII on PostScript Printers >- ---------------------------------------------------------------- I believe that other PostScript printers (in particular the Apple LaserWriter and Digital's PrintServer) will print ASCII files by having a PostScript program downloaded which "reads" the ASCII and "prints" it. Thus the printer's don't have special modes to print ASCII files, but special PostScript programs. As I remember some comment on Info-VAX, some site was running a LaserWriter from a VAX and would print ASCII files using a PostScript program which came out of a device control library! (That's a neat hack!) Frank J. Nagy Fermilab PO Box 500 MS/220 Batavia, IL 60510 (312)840-4935 ================================================================ Note 725.7 LN03-Plus vs. LN03R ScriptPrinter 7 of 15 "Bob Hassinger" 21 lines 17-AUG-1987 10:11 -< More info, please... >- ---------------------------------------------------------------- I believe that other PostScript printers (in particular the Apple LaserWriter and Digital's PrintServer) will print ASCII files by having a PostScript program downloaded which "reads" the ASCII and "prints" it. Thus the printer's don't have special modes to print ASCII files, but special PostScript programs. As I remember some comment on Info-VAX, some site was running a LaserWriter from a VAX and would print ASCII files using a PostScript program which came out of a device control library! (That's a neat hack!) VAX-64 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT If someone can find a sample of this, PLEASE make it available - perhaps here. It should be a great help for the LN03R too. I suspect there are limits on any reasonably sized program. For example straight text and program listings might not be too bad to do but it begins to be a project if you try to deal with all the escape sequences and stuff like sixels and so on that can be sent to an LN03 or other DEC printer. Even this limited capability would be very valuable however and a good starting point for a more elaborate version. Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 725.8 LN03-Plus vs. LN03R ScriptPrinter 8 of 15 "Jack Patteeuw" 11 lines 18-AUG-1987 14:10 -< A DEC Solution ! >- ---------------------------------------------------------------- Digital has announced a new software product called RETOS which does ReGIS to Sixel conversion. The part number is Q*ZE9-UZ. Pricing is **VERY** reasonable !! This means you can take your ReGIS file and convert them to Sixel and print them out on a plain old LN03 (you may need the extra RAM cartridge). Or an LA100, or LA50 or a LA75, etc. Please be aware that Sixel is **VERY** low resolution and may not be acceptable for your application. Jack Patteeuw Ford Motor Co. Electrical and Electronics Division 31630 Wyoming Livonia, MI 48150 313-323-8643 VAX-65 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 725.9 LN03-Plus vs. LN03R ScriptPrinter 9 of 15 "Jamie Hanrahan" 11 lines 18-AUG-1987 15:35 -< sixel vs. regis resolution >- ---------------------------------------------------------------- > Please be aware that Sixel is **VERY** low resolution and > may not be acceptable for your application. I don't understand -- I thought Sixel could specify every last dot on the page (at 300x300 dpi)? I should think that the resolution problems would arise from using ReGIS as a source, at 768 h x 480 v dots per screen (and the pre-VT3xx terminals only showed 240 v dots, and the Sixel output from their printer ports reflects this; the even- and odd-numbered lines are always the same). Jamie Hanrahan Simpact Associates 9210 Sky Park Court San Diego, CA 92123 619-565-1865 ================================================================ Note 725.10 LN03-Plus vs. LN03R ScriptPrinter 10 of 15 "Jack Patteeuw" 22 lines 19-AUG-1987 17:49 -< But what is a "dot" ? >- ---------------------------------------------------------------- Sixel can in fact address every dot on the page ... **FOR A DOT MATRIX PRINTER*** (ie. LA210, LA100, LA50, LA75, etc). Obviously those dots are pretty far apart, especially when you are only making one pass per line ! Nice, pretty, diagonal lines on your VT240 come out with all sort of "jaggies" when printed on a sixel printer !! When sending sixel files to a LN03, it (the LN03) simply emulates the accuracy of the LA100 !!!! (nothing near 300 dpi; I seem to recall it's something like 180 dpi). As mentioned before, BASEVIEW can convert ReGIS to TEK format which can then be printed on a LN03+. TEK 4014 mode has an "address window" of 4096 x 4096. In landscape mode (hold your sheet of paper sideways) the laser can address 3300 x 2550 (11*300 x 8.5*300) assuming no margins (ha!). You can see then VAX-66 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT that a "mapping" has to take place between the "virtual" and "physical" windows. Bottom line ... Try before you buy !!!!!! Make DEC so you some output !! Jack Patteeuw Ford Motor Co. Electrical and Electronics Division 31630 Wyoming Livonia, MI 48150 313-323-8643 ================================================================ Note 725.11 LN03-Plus vs. LN03R ScriptPrinter 11 of 15 "Bob Hassinger" 33 lines 20-AUG-1987 09:36 -< Expansion and an observation re: Regis to Sixel... >- ---------------------------------------------------------------- >> < Note 725.10 by NODE::US176598 "Jack Patteeuw" > >> -< But what is a "dot" ? >- Maybe there is a little more to the story. If, as most of the presently available Regis to Sixel programs do, you send your Regis to a terminal and read back the resulting Sixels you will get a course dot matrix, for example something like 240 by 800 from a VT240 I think. On a VT3xx I hope this will be 480 by 800. With the V2 microcode in the 200s you can get a magnified and rotated sixel array back with two by two dots for each of the raster dots but there is no increase in resolution - there are still only 800 discrete pixels across the image even though you get 1600 sixel dots. The size image these produces when sent to a printer depends on the "size of the dots" on the printer. In the case of an LN03 each sixel dot displays as two printer dots. I think you actually get a choice about this but I believe the finest resolution is a 150 by 150 array. As I recall there are a number of obscure option settings for the number of dots displayed for each sixel dot that control the pixel shape. VAX-67 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT What I am interested in is this. Note that the resolution in a VT3xx display is different from a VT240. I believe this continues to be true as you go to a VT125 and a Gigi. If the Regis to Sixel conversion is being done in software on a VAX, what is to stop you from picking any arbitrary resolution. You should be able to compute an image resolved as finely as you like, getting rid of jaggies in diagonal lines and other Regis drawing primitives. True I suspect that the locations of the primitives will most likely be limited to the 480 by 800 grid but the lines and edges could be smooth and the generated image any size you like. Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 725.13 LN03-Plus vs. LN03R ScriptPrinter 13 of 15 "Jamie Hanrahan" 59 lines 20-AUG-1987 14:15 -< re. several recent notes... >- ---------------------------------------------------------------- Nice, pretty, diagonal lines on your VT240... WHAT nice, pretty, diagonal lines?!? Is this a new rev of the VT240 that I haven't heard about? Or is yours defocused, by any chance? :-) 240 vertical dots per screen isn't "pretty" by any standards! (You know, I'll bet I could take an Atari 520ST... $900 locally, with color monitor... and program it to be a VT340 emulator, at a fraction of DEC's cost... AND I'd have a computer... but that's another story...) something like 240 by 800 from a VT240 I think 240 by 768. On a VT3xx I hope this will be 480 by 800. I hope so too. Anybody know? VAX-68 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Note that the resolution in a VT3xx display is different from a VT240. I believe this continues to be true as you go to a VT125 and a GIGI. No. VT125 and VT240/241 are identical in resolution (there are some differences in the display hardware, though). GIGI's single-bit graphics bitplane is essentially the same resolution but has, if I recall correctly, 10 more horizontal scan lines (and so can display 25 lines of text). Its "display attributes memory", on the other hand, provides four bits of display info (three for color select and one for blink) for each set of 12 horizontal pixels, for a resolution of only 64 h by 250 v. Truly a bizarre device. You should be able to compute an image resolved as finely as you like, getting rid of jaggies in diagonal lines and other Regis drawing primitives. True I suspect that the locations of the primitives will most likely be limited to the 480 by 800 grid but the lines and edges could be smooth and the generated image any size you like. There's really no reason to limit to 480x768. With the old ReGIS Graphics Editor I can make images that cover four screens, specifying ~1500x1000 dots. Can't do much with them after that, of course... but the 480x768 limit is imposed by the terminals, not ReGIS definitions. Jamie Hanrahan Simpact Associates 9210 Sky Park Court San Diego, CA 92123 619-565-1865 VAX-69 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 726.0 Software DES Implementation No replies "Offline Submission" 12 lines 11-AUG-1987 21:59 ---------------------------------------------------------------- Does anyone have a public domain implementation of DES (Data Encryption Standard) or similar public key encryption software. It does not have to be fast -- I only need to encrypt/decrypt 64 bytes of data. Don Rubin S.E.T. Incorporated 9703 Forest Glen Court Silver Spring, MD 20910-1121 301-588-8010 July 24, 1987 ================================================================ Note 727.0 RUNOFF .REQUIRE limitation?? 6 replies "Kevin Angley" 16 lines 12-AUG-1987 14:25 ---------------------------------------------------------------- I have reported what appears to be a bug/restriction with Runoff: There seems to be a limit on the total number of .REQUIRES that can be contained in a source stream (Note - this shouldn't have anything to do with nesting or with # of open files, as the .REQUIRES are sequential). Example: A file contains 150 lines each containing .REQUIRE "TEST1.RNO". TEST1.RNO is one line of text. After the first approx. 120 lines, it will fail to include the file because of an "unexpected operating system error". Doing a SET WATCH FILE seems to indicate that there is no access attempted immediately prior to getting the message. Does anyone have a clue? Kevin Angley 3301 Terminal Drive Raleigh, NC 27604 (919) 890-1416 VAX-70 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 727.1 RUNOFF .REQUIRE limitation?? 1 of 6 "Bob Hassinger" 4 lines 13-AUG-1987 18:07 -< File close bug? >- ---------------------------------------------------------------- Does this have something to do with a bug that fails to close files after they have been used? I seem to recall something like that quite a while back and I think it may have been in DSR. Are you using a current version? Bob Hassinger Liberty Mutual Research Center 71 Frankland Road Hopkinton, MA 01748 617-435-9061 ================================================================ Note 727.2 RUNOFF .REQUIRE limitation?? 2 of 6 "Bruce Bowler" 2 lines 14-AUG-1987 15:33 -< see note 650.* >- ---------------------------------------------------------------- See the chain of reply's for note 650 [ Pageswapper Volume 8 Number 12 ], it talks some more about this problem. Bruce Bowler General Electric 1 River Road Bldg 2 Room 609 Schenectady, NY 12345 VAX-71 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 727.3 RUNOFF .REQUIRE limitation?? 3 of 6 "Chris Erskine" 11 lines 18-AUG-1987 11:43 -< One fix for bug >- ---------------------------------------------------------------- Does this have something to do with a bug that fails to close files after they have been used? I seem to recall something like that quite a while back and I think it may have been in DSR. Are you using a current version? This does sound like the file close bug. DSR does not deassign the channel that is assigned for the file from a require. Even though the file is closed, the channel stays around and after about 128 .REQUIRE's, you die. There is a parameter which I think is in SYSGEN which can be used to increase this number but I do not remember it right now. Chris Erskine 6001 Adams Rd. Bloomfield Hills, MI 48013 (313) 258-4049 ================================================================ Note 727.4 RUNOFF .REQUIRE limitation?? 4 of 6 "Kevin Angley" 4 lines 18-AUG-1987 17:15 -< More info >- ---------------------------------------------------------------- Just to confirm .. we are using 4.5 version of RUNOFF. Doesn't deassign the channel ... any clues as to what may need to be increased to work-around this problem? Kevin Angley 3301 Terminal Drive Raleigh, NC 27604 (919) 890-1416 VAX-72 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 727.5 RUNOFF .REQUIRE limitation?? 5 of 6 "Chris Erskine" 3 lines 19-AUG-1987 09:11 -< Change in UAF >- ---------------------------------------------------------------- Try increasing FILLM in UAF. This also uses additional BYTLM so this value may have to be increased. Chris Erskine 6001 Adams Rd. Bloomfield Hills, MI 48013 (313) 258-4049 ================================================================ Note 727.6 RUNOFF .REQUIRE limitation?? 6 of 6 "John Burnet" 7 lines 19-AUG-1987 13:01 -< Try upping CHANNELCNT in sysgen >- ---------------------------------------------------------------- Try increasing FILLM in UAF. This also uses additional BYTLM so this value may have to be increased. I don't think that will do it; I think you'll have to increase the sysgen parameter CHANNELCNT (default is 127, maximum is 4095 or something huge like that). It isn't a dynamic parameter, so you'll have to reboot after changing it. John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 VAX-73 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 728.0 Dump files 5 replies "Howard W. Pinsley" 26 lines 12-AUG-1987 17:49 ---------------------------------------------------------------- I have a 4 node cluster that is utilizing a shared RA-81 system pack. The four systems are: 8700 - 80mb 8650 - 64mb 785 - 64mb 785 - 64mb We have disabled dumping due to the amount of disk space (approx 480,000 blocks) that would be required on our system pack. I would like to enable system dumps on all my systems with the minimum amount of pain. The only solution I know of is to use more than 1 system pack. I would rather not resort to that. Is there a way to allocate only one file that is big enough to accommodate the largest dump for all the systems. Also, I heard rumors that it might be possible to use a secondary page file (on another disk) as a dump file. Any insight would be appreciated. P.S. I have been told (via the ARIS bbs) that it is possible to have a common dump file (i.e. sys$common:[sysexe]sysdump.dmp) as long is it is big enough to accommodate the largest memory requirement. Does anyone know about this technique? Any why wouldn't DEC mention this in the sysmgr guide or the guide to VaxClusters? Howard W. Pinsley 145 West 79th. Apt 11D New York, NY 10024 VAX-74 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 728.1 Dump files 1 of 5 "Bruce Bowler" 11 lines 14-AUG-1987 15:44 -< Should work, but... >- ---------------------------------------------------------------- I ran into somebody at the DECUS 2 trips ago who did this (I don't remember who, it was that last day) and he said things worked just fine with one large dump file, EXCEPT when multiple machines died at the same time, if that happens only the dump of the last system to "bit the big one" will be available for perusal. I believe he also said that all of the systems sharing a dump file, must have the same amount of memory. If this is the case, you would need 2 dump files, one for your 80Mb machine and one for your 64Mb machines. I can't say for sure, but my guess as to why DEC doesn't tell us that we can do this is because they want to sell us more disks. Bruce Bowler General Electric 1 River Road Bldg 2 Room 609 Schenectady, NY 12345 ================================================================ Note 728.3 Dump files 3 of 5 "Frank J. Nagy" 20 lines 15-AUG-1987 11:24 -< SYS$COMMON:[SYSEXE]SYSDUMP is OK! >- ---------------------------------------------------------------- YES! It is possible to put SYSDUMP.DMP into SYS$COMMON:[SYSEXE] and have a single dump file for all your systems. At Fermilab, my group's two LAVCs have done this and the BIG VAXCluster has also done this. No, all the machines do not have to have the same memory size and only one SYSDUMP.DMP file is needed for all (make it the memory size of your largest machine + 4 blocks). If 2 or more machines die AT THE SAME time, then you will end up with a meaningless dump file contents since the writes to the file are totally uncoordinated and who knows who wrote what, where in the file and when - this is only the case where a single bugcheck crashes all or most of the cluster and 2 or more machines get to the "writing dump file" stage at very nearly the VAX-75 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT same times. NOTE: If your dump file is not large enough to hold all of a system's memory, then nothing bad will happen (i.e., your disk does not get overwritten), VMS will only write the portion of memory which will fit in the file. This might cause the dump to be useless for debugging a problem but is not otherwise a problem. Frank J. Nagy Fermilab PO Box 500 MS/220 Batavia, IL 60510 (312)840-4935 ================================================================ Note 728.4 Dump files 4 of 5 "Lee K. Gleason" 29 lines 18-AUG-1987 00:27 -< The Space Saver Spare >- ---------------------------------------------------------------- I wrote the note on the ARIS board about using a single common dump file for multiple VAXen on a cluster. At one of our sites, we have a 48 Mb 8600, a 16 Mb 785, and an 8Mb 8200 using the same dump in SYS$COMMON:[SYSEXE], sized for the 48Mb+4 pages required by the 8600, and it works for all three of them. Reasonably enough, this is a problem if multiple VAXen crash at once - but, hell, man, if that happens, you don't need a crash dump to tell you you're in trouble. If you use a file that is too small, the dump file will definitely be useless for debugging - high physical memory is where the system and global page tables, and the balance slots are loaded. Desperation drove us to look for this space saving solution. We don't have a block to spare here. Before trying the SYS$COMMON approach, we had a slightly more exotic solution - we entered aliases for the same file in all three SYS$SPECIFIC:[SYSEXE] directories, with SET FILE/ENTER. This worked, but it was a real nuisance dealing with a file with multiple directory entries. VAX-76 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT DEC probably doesn't mind the disc drive sales caused by dump files, page files, and swap files... Lee K. Gleason Control-G Consultants 2416 Branard #D Houston TX 77098 713/528-1859 ================================================================ Note 728.5 Dump files 5 of 5 "Kevin Angley" 6 lines 18-AUG-1987 17:09 -< I learned it at DECUS >- ---------------------------------------------------------------- We also use a common sysdump file with no problems. At system startup, we check to see if the dumpfile contains a crash dump. If so, we copy it to another file. That way, as long as two systems don't crash within 10-15 minutes of each other, all dumps are saved. Another solution (tongue in cheek): Buy less memory. Kevin Angley 3301 Terminal Drive Raleigh, NC 27604 (919) 890-1416 ================================================================ Note 729.0 Strange behaviour of DEFINE and SHOW LOGICAL2 replies "Offline Submission" 77 lines 14-AUG-1987 18:00 ---------------------------------------------------------------- We recently had a problem (with All-in-1) which was traced back to a logical symbol being lower case. The logical should have been the equivalent of: DEFINE KOA$TERMINAL_VTAnn "something" but was actually the equivalent of: DEFINE koa$terminal_VTAnn "something" VAX-77 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Of course, when the software did a translation, it somehow got the wrong (or no) answer. We still don't understand exactly what is going on - but the following illustrates the problem. Is this an undocumented feature, or a bug?? DEFINE "a" "lowercase" SHOW LOGICAL A "a" = "lowercase" (LNM$PROCESS_TABLE) SHOW LOGICAL a "a" = "lowercase" (LNM$PROCESS_TABLE) SHOW LOGICAL "a" "a" = "lowercase" (LNM$PROCESS_TABLE) SHOW LOGICAL "A" "a" = "lowercase" (LNM$PROCESS_TABLE) (Well, that's the expected behaviour. Now .....) DEFINE "a" "UPPERCASE" (Notice there is no message that previous definition was superseded) SHOW LOGICAL A "A" = "UPPERCASE" (LNM$PROCESS_TABLE) SHOW LOGICAL a "A" = "UPPERCASE" (LNM$PROCESS_TABLE) SHOW LOGICAL "a" "A" = "UPPERCASE" (LNM$PROCESS_TABLE) SHOW LOGICAL "A" "A" = "UPPERCASE" (LNM$PROCESS_TABLE) (But it looks like the definition really was superseded. Now, just for fun, let's look at everything beginning with A...) SHOW LOGICAL "a*" (LNM$PROCESS_TABLE) "a" = "lowercase" [ Why "a*" shows this and "a" doesn't is a mystery] (LNM$JOB_803CFC90) (LNM$GROUP_000001) (LNM$SYSTEM_TABLE) VAX-78 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT (Well, that's really interesting. Let's get rid of the definition...) DEASSIGN A SHOW LOGICAL A "a" = "lowercase" (LNM$PROCESS_TABLE) Hm...... :-} What's going on here? Just for grins :-) you might do a SHOW LOGICAL/ALL/OUTPUT=filename on your system and look for lowercase stuff on the left hand side. Dale E. Coy 717 Meadow Lane Los Alamos (WR), NM 87544 July 29, 1987 ================================================================ Note 729.1 Strange behaviour of DEFINE and SHOW LOGICAL 1 of 2 "Chris Erskine" 12 lines 18-AUG-1987 11:57 -< Solution and additional fun >- ---------------------------------------------------------------- Since DCL translates anything that is not in quotes to uppercase, most of the time this will not be seen. The SYS$TRNLNM service has a flag to translate in a case blind mode. From what you are saying, the SHOW LOGICAL command is doing it's translation in blind mode. Once it finds a match for the logical you are asking for, it does not try to translate any other logicals. The system service SYS$CRELNM can also cause this and some other problems. If you create the logical 'A ', it is stored in the logical table with the space in it and can make it difficult to use or delete it. Chris Erskine 6001 Adams Rd. Bloomfield Hills, MI 48013 (313) 258-4049 VAX-79 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 729.2 Strange behaviour of DEFINE and SHOW LOGICAL 2 of 2 "Dale E. Coy (505) 667-3270" 14 lines 21-AUG-1987 19:09 -< Supplemental info >- ---------------------------------------------------------------- Thanks for typing all of that in - and now I'm online. The real fun (?) in all of this is the difference between SHOW LOGICAL "a" and SHOW LOGICAL "a*" One would think that the "a" would be a subset of "a*", but it ain't so. The explanation HAS to be wrapped up in case_blind/case_sensitive, but it sure doesn't make sense. Credit goes to Atlanta for FIXING our problem, even when they didn't know what caused it (I found the cause after the fix was made). DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 730.0 TPU Editor Problem 4 replies "Offline Submission" 51 lines 14-AUG-1987 18:24 ---------------------------------------------------------------- Hello Experts out there, as a newcomer to VMS I am bound to get into some unsolvible troubles. I wonder if anyone could offer some help? I have a MicroVAX-II with TK50, RD54 and DHV11 and I am having troubles starting the TPU editor. No matter how I define and/or initiate it, it always comes back with the normal key pad editor. Here is my relevant part of SYSTARTUP.COM: $ RUN SYS$SYSTEM:INSTALL $!sys$system:analimdmp /priv=(cmexec,cmkrnl) sys$system:mail /open/header/priv=(sysprv,oper,world,netmbx) sys$system:phone /priv=(netmbx,oper,prmmbx,world,sysnam) $!sys$system:request /priv=(tmpmbx) VAX-80 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT $!sys$system:rtpad /priv=(tmpmbx) sys$system:submit /open/header/priv=(tmpmbx) sys$system:tpu /open/header sys$system:tpu$cctshr /open/header sys$system:tpushr /open/header sys$system:tpumsg /open/header $! $! WPS+ Files needed $! $sys$share:vmsrtl /open/header/shared $sys$share:mthrtl /open/header/shared $sys$share:lbrrtl /open/header/shared $sys$share:scrrtl /open/header/shared $sys$share:convrtl /open/header/shared $ $! RUN SYS$SYSTEM:INSTALL $!sys$system:copy /open/header/shared $!sys$system:delete /open/header/shared $!sys$system:edt /open/header/shared $!sys$system:rename /open/header/shared $ When I type "EDIT/TPU FILENAME" I get "extra characters ignored" and the normal Key Pad Editor appears. Any suggestions will be very much appreciated. Rudi W. Stange c/o Digital Equipment GmbH Freischuetz Str. 91 D8000 Munich 81 FRG August 3, 1987 ================================================================ Note 730.2 TPU Editor Problem 2 of 4 "Michael R. Pizolato" 12 lines 21-AUG-1987 14:07 -< A quick suggestion >- ---------------------------------------------------------------- If that is an exact copy of the relevant parts of your SYSTARTUP.COM, then I would suggest changing all the "$!" strings to "!" strings. INSTALL might be exiting without doing anything at all, since when it encounters the first "$!" it probably thinks it has reached the end of the input deck. Also, remove all the "$" in front of file names, for the same reason. VAX-81 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT I am not familiar with microVMS - but this might help. Michael R. Pizolato AT&T Technology Systems Dept. 323610 555 Union Blvd. Allentown, PA 18103 215/439-5500 ================================================================ Note 730.3 TPU Editor Problem 3 of 4 "Dale E. Coy (505) 667-3270" 6 lines 21-AUG-1987 23:21 -< I'd bet on EDIT being redefined >- ---------------------------------------------------------------- TPU is no different on microVMS, of course. We've had several users try to redefine EDIT, with extremely funny results. I'd bet on it. If SHOW SYMBOL EDIT or SHOW LOGICAL EDIT show anything at all, find out why. DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 730.4 TPU Editor Problem 4 of 4 "John P. Hoekstra" 10 lines 29-AUG-1987 14:00 -< LOGIN.COM is the culprit >- ---------------------------------------------------------------- Ah yes .3 is exactly correct. The MicroVMS LOGIN.COM file in the sysmgr directory contains the line: E*DIT :== EDIT/EDT/COMMAND=SYS$LOGIN:EDTINI.EDT VAX-82 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Which, of course, ends up adding the /TPU qualifier to the end of the EDIT/EDT string and giving an error. You can remove this line from LOGIN.COM if you desire to use TPU, or any other editor for that matter. John P. Hoekstra Macatawa Computer Services, Inc. 1 West Fifth Street Holland, MI 49423-2908 (616)392-6941 ================================================================ Note 732.0 VMS JOBSCHEDULER NEEDS IMPROVEMENT 3 replies "GREG P. SCHULZ" 24 lines 15-AUG-1987 10:36 ---------------------------------------------------------------- We are a large IBM shop with VAX's and other minis. Our users are new to the VMS environment and have expressed concerns over the lack of "basic" facilities such as a job scheduler. Granted the users could write some DCL code to perform such things as checking for device allocation or if files exist prior to execution. They could also use the job controller to schedule execution. of jobs. Since our users have had utilities available in the past for performing such tasks, we need to find or develop something. VAX hardware and layered products may be catching up to or surpassing IBM, but what about some of the basic utilities. Granted V4.x gave us new and improved printing capabilities, but how about support for batch processing improvements. If VAX/VMS is supposed to be viable commercial O.S. for use in commercial corporate computing, where are the tools to support such applications? VMS is lacking in the areas of tape management, and job scheduling to name a few. We are looking for some sort of job scheduler for the VAX that will provide more then an easy to use "submit" facility. What VMS needs is a form of the IBMUCC7 job scheduler. Is there anyone out there that can help? DEC sure has not done a heck of lot... P.S. We have found a solution to the tape problem... GREG P. SCHULZ BURLINGTON NORTHERN RR ISS LOC 3 VAX-83 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT 176 E 5TH STT ST. PAUL, MN 55164 612 298-7344 ================================================================ Note 732.1 VMS JOBSCHEDULER NEEDS IMPROVEMENT 1 of 3 "Jack Patteeuw" 10 lines 17-AUG-1987 11:06 -< What does IBM do ? >- ---------------------------------------------------------------- Greg, Yes I agree that VMS **DESPERATELY** needs a tape handling/ archive system. Furthermore, I am disturbed to hear that it won't be in V5.0 But my question to you is "what kind's of thing does your IBM job scheduler do that the present VMS queue manger can't do ?" I am asking because I believe that this area also needs work and I'd like to know how other O.S. do it. Jack Patteeuw Ford Motor Co. Electrical and Electronics Division 31630 Wyoming Livonia, MI 48150 313-323-8643 ================================================================ Note 732.2 VMS JOBSCHEDULER NEEDS IMPROVEMENT 2 of 3 "Dale E. Coy (505) 667-3270" 11 lines 21-AUG-1987 23:35 -< //COMMENT >- ---------------------------------------------------------------- I'd be interested, too. Back when I was writing JCL, users were expected to do it. Later, you could get stuff from the users group, and the system staff wrote big JCL so the poor users could write simple JCL. I find it hard to conceive of JCL (or VMS Submit) which is smart enough to check for device allocation, file existence, etc. without ANY action by the user. Is this Artificial Intelligence? VAX-84 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT I presume that the IBM stuff you quote was written by IBM and supplied with the operating system. Can you tell us more about what it does? DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 732.3 VMS JOBSCHEDULER NEEDS IMPROVEMENT 3 of 3 "John Osudar" 27 lines 25-AUG-1987 16:08 -< what's necessary ain't easy >- ---------------------------------------------------------------- There is definitely plenty of room for improvement in VMS batch job scheduling; however, it may require a major overhaul to the way things work in VMS in order for DEC to implement something really useful in this area. Consider, for example: it would be useful to specify that a batch job requires devices AAA and BBB in order to be eligible to run. The straightforward approach is to check those devices periodically, and if they are ever both available at the same time, start the job (and hope that nobody else grabs them while the job is starting up!) A more useful (more correct?) approach is to centralize device (and other resource) allocation through a common "resource manager". This piece of software would track allocations, requests, and job "requirements", and everyone would be required to go through it to allocate such resources. Then, each job could get a resource "tentatively allocated" to it -- and if a job got all the resources it needs to run, that job would be allowed to start. (This could even be extended to work piecemeal on "job steps".) Other jobs could preempt tentative allocations if that would enable them to run. All kinds of other bells and whistles can be added onto this -- priority-based preferential allocation, for example -- but the key factor is that the central "resource manager" is REQUIRED in such a system. VMS has the tools to implement such a beast, but I have heard of no plans by DEC to do so. Perhaps if enough of us want it, and if we can define it clearly enough, we can get it on the SIR list and get DEC to think about it. John Osudar Argonne National Laboratory 9700 S. Cass Ave. VAX-85 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Bldg. 205 A-051 Argonne, IL 60439-4837 (312) 972-7505 ================================================================ Note 733.0 Strange DMB-32 problems 3 replies "John Burnet" 60 lines 20-AUG-1987 11:08 ---------------------------------------------------------------- We're experiencing some very strange problems with the two DMB-32s on our VAX 8200. We have several 1200-baud modems (Hayes & Vadic) connected to ports on both DMBs; the modems are used for both dial-in access to the system and dial-out access to other systems. We also have a few terminals and printers connected to DMB ports. At times the DMB (or its driver) will appear to assert carrier on a port when no carrier is present; at the same time, a process reading the port will see a continuous stream of garbage. This goes on for long periods of time (it happened for 8 hours once before "carrier" was dropped), and can usually be cleared temporarily by rebooting the system. At other times the DMB (or the driver) will drop carrier on a port for no apparent reason, cutting off any session in progress. At first we thought that this might be due to phone line noise or attenuation problems, but that doesn't appear to be the case. (For one thing, we've managed to get the DMB to drop DCD and/or DSR on a port while a terminal was hard-wired to the port with a cable.) Another symptom that shows up occasionally is that a port will simply become wedged and refuse to respond--doing a set host/dte to a port and trying to give commands to the modem connected to the port will elicit no response, and (to add insult to injury) control-backslash won't work then and the process will have to be forcex'd from another process. (Interestingly enough, when the forcex is done, the characters which were typed during the set host/dte show up in the terminal typeahead buffer and echo on the terminal. Also, ^\ works if no other character is typed [sent to the port] between the set host/dte command and the ^\.) VAX-86 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT There's nothing wrong with the SET TERM port characteristics, as far as we can tell. We might experience these problems with a port, then reboot and have no further problems with that port for a while even though the SET TERM settings are exactly the same. The settings do include /MODEM/HANGUP. DEC Field Circus has been out to have a look and has swapped DMB-32 controllers (including a swap to the next higher rev level) with no apparent result. They checked out the patch panel and the wiring and found nothing wrong. Now they're trying to tell us that "it's a software problem." Great, but with what software? The driver? F. C. checked with Colorado Springs and found no known problems with YIDRIVER that could be causing this. Does anyone have any ideas as to where the problem might lie? We've heard of the YIDRIVER even/odd parity confusion problem, but haven't heard of any other driver bugs that might account for the symptoms we're seeing. John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 ================================================================ Note 733.1 Strange DMB-32 problems 1 of 3 "Bill Mayhew" 19 lines 20-AUG-1987 17:38 -< I'm confused. >- ---------------------------------------------------------------- I'm not quite sure I understand. Carrier-detect can only be asserted *by* a modem *to* a comm device; the DMB should be detecting the carrier-detect signal, not asserting it. I have seen situations where a modem will suddenly think it has a carrier and start chattering away to the host. Some of these have turned out to be flaky modems, others have turned out to be very flaky phone lines. Haven't seen any that were comm-port problems, but that doesn't mean it can't happen. What are you seeing in the lights on these modems (assuming they have lights?) VAX-87 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT I have heard of modems that would assert carrier-detect back to the host in response to the host raising some other control signal (DTR?). This is done so that the host can think it sees a carrier, and thereby the host will deign to allow data (e.g. an autodial command) to pass from the host to the modem when software and/or hardware would otherwise prohibit such passage in the absence of a carrier-detect signal. Have never touched such a beast, however. Might this be your situation? Bill Mayhew Village Systems Workshop Inc PO Box 642 Natick MA 01760 617-237-0238 ================================================================ Note 733.2 Strange DMB-32 problems 2 of 3 "Larry Kilgallen" 6 lines 20-AUG-1987 21:03 -< TT2$V_XON for one and all >- ---------------------------------------------------------------- The situation where $FORCEX causes VMS to spill its guts sounds like the line might be XOFF'd. Various cleanup code in the terminal driver clears XOFF. You can do so too, with the TT2$V_XON bit as part of an IO$_SETMODE $QIO to the device (if you can get it to accept your $QIOs at all. None of this explains how the line got into that state. Larry Kilgallen Box 81, MIT Station Cambridge, MA 02139-0901 ================================================================ Note 733.3 Strange DMB-32 problems 3 of 3 "John Burnet" 31 lines 22-AUG-1987 13:32 -< Good ideas! Also some clarifications... >- ---------------------------------------------------------------- Part of my description of the problem in .0 was a little misleading... I finally realized (it should have been obvious) that a terminal connected directly to a port shouldn't have /MODEM enabled. (For test purposes, I had unplugged a modem from the DMB and plugged a terminal into that port without setting the port /NOMODEM.) However, the problem remains, since all of our actual modem lines *do* have /MODEM set, and it's on VAX-88 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT those lines that we've been seeing the problems. Re: .1, what lights am I seeing on the modem... good question! Most of the time, the problem shows up in processes that run in batch and as detached processes, usually at odd hours of the day when no one's around. (We do data collection over the phone from some remote process-control sites, and also have a connection to Western Union's Easylink; we've seen the problem with both of those connections.) I'll have to watch carefully for the next time the problem occurs (if it does--I don't know whether to hope it will or not), and check the modem for CD, RXD, TXD, etc. The modem option switches are set such that CD is *not* asserted to the host unless carrier is actually present. Thanks, Larry, for the xon suggestion... I knew about SET TERM/XON but didn't stop to think that that might be the problem when getting no response during a set host/dte... I'll have to try that. Our LXY12 on the DMB printer port had been spewing out garbage, too. That appeared to be fixed after the controller swap, but the async problems have continued. It's beginning to look more and more like those problems *are* due to flaky cables or modems... we'll have to "break out" the breakout box and check everything. John Burnet P. O. Box 1838 Evanston, IL 60204 (312) 272-6520 x2118 ================================================================ Note 734.0 Transfer files to another company (batch Kermit?). 2 "Michael R. Pizolato" 16 lines 21-AUG-1987 14:24 ---------------------------------------------------------------- We need to transfer files from our VAX to another company's VAX. Both run VMS, but they are not on the same network, so asych DDCMP is no good because of possible address conflicts, also because our network management won't allow another company's machine to be defined on our network. VAX-89 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT More info: we will have to use 1200 baud modems. We have DEC DF112's; I don't know what they have. Mag tape via the mail is no go. All transfers are from us to them, and will be initiated by us. Both machines have KERMIT. We would like to do this regularly (maybe once a day, maybe less often), in batch. Any suggestions? They're breathing down my neck for this... :-) Michael R. Pizolato AT&T Technology Systems Dept. 323610 555 Union Blvd. Allentown, PA 18103 215/439-5500 ================================================================ Note 734.1 Transfer files to another company (batch Kermit?). 1 "Jamie Hanrahan" 4 lines 21-AUG-1987 15:37 -< Try C-Kermit >- ---------------------------------------------------------------- C-Kermit, which is on the DECUS SIG tapes along with VMS Kermit, runs under VMS and supports autodial modems, login script negotiation, etc. I've used it for just that sort of thing. Let me know how it works out. Jamie Hanrahan Simpact Associates 9210 Sky Park Court San Diego, CA 92123 619-565-1865 VAX-90 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 734.2 Transfer files to another company (batch Kermit?). 2 "M. Erik Husby" 12 lines 26-AUG-1987 09:21 -< Batch file transfers with VAXNET >- ---------------------------------------------------------------- Another alternative is VAXNET, which also supports autodial modems, batch procedures, and several transfer protocols. We use it to send VAX MAIL between our branch offices and main office as well as talking to Western Union's TELEX system. All the transfers are done through batch jobs that run at periodic intervals or on demand. VAXNET is available on the DECUS tapes. For the security minded, it has ways to limit and track the phone usage. M. Erik Husby Project Software & Development 14 Story St. Cambridge, MA. 02138 (617)-661-1666 ================================================================ Note 735.0 Bar Codes 2 replies "Howard W. Pinsley" 24 lines 24-AUG-1987 23:57 ---------------------------------------------------------------- I am developing an application that, in many respects, resembles a library system (in the sense that items are checked in and out). The system is designed to produce labels to facilitate the filing process. I would like to print a bar codes on the labels to expedite the sign in/out processes. My preliminary inquiries have yielded the following: A VT2XB is a small box with a light pen that sits near your keyboard. It performs the read function by feeding translated ascii characters to your terminal just as your keyboard normally would. VAX-91 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT An LG02 printer is a high speed/high quality printer for bar coding. It will intermix ascii and bar codes. It is pretty expensive and not practical for my environment. I would like several desk top printers for several sites. Someone THINKS they heard that either the LA210 or LA75 do bar codes. I have not been able to confirm this. Can anyone supplement or correct this information. In addition, any general advice on implementing this functionality would be welcomed. Howard W. Pinsley 145 West 79th. Apt 11D New York, NY 10024 ================================================================ Note 735.1 Bar Codes 1 of 2 "MICHAEL GRATTAN" 9 lines 27-AUG-1987 07:34 -< A thought... >- ---------------------------------------------------------------- According to the VAX System & Options Catalog (April - June 1987) both the LA75 and the LA210 support full bit mapped graphics. (The LA75 says 180 by 144 dots per inch resolution.) This sounds to me like they 'should' be able to handle bar codes, but Dec doesn't come out and say so. There are various third party solutions which might not be as pricey as the LG which you might want to consider. MICHAEL GRATTAN FAIRHAVEN CORP. 358 BELLEVILLE AVE. NEW BEDFORD, MA. 02742 617-993-9981 EXT 106 VAX-92 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 735.2 Bar Codes 2 of 2 "Dale E. Coy (505) 667-3270" 21 lines 27-AUG-1987 12:30 -< As usual, it depends.... >- ---------------------------------------------------------------- Our experience says you MAY be better off with a third party solution. Believe it or not, this problem has much in common with the "quick and dirty" software question. If you're doing a "small" application, or a prototype, or something just for yourself, or trying to learn something and/or your time is "worth" less than anything else, then by all means hack at it. A dot-matrix printer will give you minimal bar-code quality, but it does work. Any laser printer will do better. LN-03 (get the programmers guide) is OK. HP Laserjet has a plug-in cartridge for bar codes. If you're doing a "real" application, then go with hardware from somebody who really understands bar codes. There are lots of standards, etc... The production applications used here all are based on 3rd party printers (some stand-alone and some RS-232 interfaced) and self-contained readers which dump their data to computers (on-line-real-time or batched). There are lots of issues: readability of labels (human and machine), ability to stand up for a long time in a non-office environment, etc... DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 VAX-93 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 736.0 DECnet-DOS applications No replies "M. Erik Husby" 26 lines 26-AUG-1987 09:35 ---------------------------------------------------------------- I am presently trying to convert the DECNET-DOS routines from the two-segment memory model to the large memory model using the Microsoft Version 4 C compiler. I am doing this because I am building a PC application that will communicate with a server on a VAX. I am having problems with random failures -- it seems someone is writing over memory somewhere. Questions: 1. Has anyone else already done this? If so could you contact me to talk about your experiences. 2. I am doing my development on VAXmate but will need to have the code run on ordinary PCs. Any thing to be aware of here. 3. Has anyone else developed a PC application using DECNET-DOS that communicates with a server on a VAX? I am trying to develop some error recovery and reconnect techniques that may be useful. I think I can make the network server reconnectable similar to disconnectable terminals. M. Erik Husby Project Software & Development 14 Story St. Cambridge, MA. 02138 (617)-661-1666 ================================================================ Note 737.0 Magic ACE bug No replies "Kevin Angley" 42 lines 26-AUG-1987 10:15 ---------------------------------------------------------------- This is somewhat of a rehash of notes 676.*, but its original title may have scared away non-CMS users when in fact it is a VMS problem. It is the issue of the "Magic ACE". Consider the following: $ show process/privilege TEST dynamic, resource VAX-94 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT $ directory [-]test.dir/security TEST.DIR;1 1 25-AUG-1987 17:06 TEST (RWE,RWE,RWE,R) (DEFAULT_PROTECTION,SYSTEM:RWE,OWNER:,GROUP:,WORLD:) (IDENTIFIER=TEST,OPTIONS=DEFAULT, ACCESS=READ+WRITE+EXECUTE+DELETE+CONTROL) (IDENTIFIER=TEST,ACCESS=READ+WRITE+EXECUTE+DELETE+CONTROL) $ create foo.bar $ directory/security foo.bar FOO.BAR;1 0 26-AUG-1987 09:12 TEST (RWE,,,) (IDENTIFIER=[FARKLE],OPTIONS=NOPROPAGATE,ACCESS=CONTROL) (IDENTIFIER=TEST,ACCESS=READ+WRITE+EXECUTE+DELETE+CONTROL) What is happening here is that the automatic magic ACE is added to the top of the ACL to give the "owner" rights identical to those granted to him by the UIC owner field plus control. I am guessing that the developers figure this could not cause a problem since the ACE provides only that access which would have been his through UIC based protection. Wrong. This only works if Access granted by UIC >= Access granted by ACL (for that "owner"). If this is not the case, the magic ACE blocks the "owner" (who isn't the owner anyway - the resource identifier is). In the example above, Farkle is not able to read the file any more even though the ACL would have otherwise allowed him to, because of the magic ACE bug. Most of the answers to this problem under Note 676.* say to give the UIC-based owner RWED so that the magic ACE will be RWEDC. But this misses the point. I do not want the former "owner" of the file to have the special ACE, because the application needs to control access via the TEST identifier only, without worrying about the spurious extras that tag along. Any wisdom on this? Kevin Angley 3301 Terminal Drive Raleigh, NC 27604 (919) 890-1416 VAX-95 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT ================================================================ Note 738.0 WANTED: VAXNET version 12 No replies "BRIAN HAJDER" 9 lines 26-AUG-1987 13:17 ---------------------------------------------------------------- I have recently gotten VAXNET, version 11, going on our Microvax to let users dial out to remote sites on a Hayes Smartmodem 2400. This has been working well but I would be interested in getting version 12 of VAXNET. I am new to using PAGESWAPPER I/O and DECUS in general, but would appreciate finding a source to download from. We use VMS Kermit a lot, but I think I could also use xmodem or sndrcv. BRIAN HAJDER DESPATCH INDUSTRIES, INC. PO BOX 1320 MINNEAPOLIS MN 55440 612-331-1873 ================================================================ Note 739.0 When should RSM and/or VCS be used? 1 reply "Dale E. Coy (505) 667-3270" 20 lines 29-AUG-1987 16:11 ---------------------------------------------------------------- There are a least 3 DEC products which facilitate management of modest quantities of VAXen/microVAXen: LAVC (Local Area VAXClusters) RSM (Remote System Manager) VAXCluster Console System (hardware and software) Question: What combinations of these products are useful? I understand that each product has it purpose, limitations, etc., but ... I'd like to hear from someone with real experience with RSM and/or VCS. Postulate a modest number (say 16) of microVAXen, and maybe a miniVAX, in an environment with DECnet running over Ethernet. LAVC is an obvious start (or maybe not so obvious). Would one also benefit from some of the features of RSM in this case? How about VCS? VAX-96 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT Maybe a better question is: Compare and contrast the benefits of RSM and VCS. Is there some experience out there? DALE E. COY LOS ALAMOS NATIONAL LAB PO BOX 1663, MS J957 LOS ALAMOS, NM 87545 505-667-3270 ================================================================ Note 739.1 When should RSM and/or VCS be used? 1 of 1 "Frank J. Nagy" 17 lines 30-AUG-1987 13:15 -< Love LAVC and VCC; Yet to use RSM >- ---------------------------------------------------------------- We use all three products. We have two LAVCs (2 and 3 MicroVAX nodes each). We have RSM but have not yet used it (being too busy with other things). We have a VAXCluster Console system. It replaces the LA120 consoles for the 3-node LAVC and 3 stand-alone (i.e., non-LAVC) MicroVAXes. VCC has already proved its usefulness to us; the VCC node, HYDRA, is connected to the network and we use it to check on the systems which are in another building. Our hardware person uses HYDRA to watch the maintenance people running diagnostics. LAVC is just great. We have RSM licenses but haven't used it yet. That will change are we are planning to buy more MicroVAXes for stand-alone systems. Unlike the three I mentioned above (which each have TK50s), the new systems will have an RD54 but no load media at all. We are going to depend upon RSM to down-line load the system for the initial setup and use RSM to backup and maintain (install updates) in these systems. So, in a few months... Frank J. Nagy Fermilab PO Box 500 MS/220 Batavia, IL 60510 (312)840-4935 * VAX-97 PAGESWAPPER - October 1987 - Volume 9 Number 3 INPUT/OUTPUT * VAX-98 PAGESWAPPER - October 1987 - Volume 9 Number 3 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-99 PAGESWAPPER - October 1987 - Volume 9 Number 3 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-100 PAGESWAPPER - October 1987 - Volume 9 Number 3 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-101 PAGESWAPPER - October 1987 - Volume 9 Number 3 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-102 PAGESWAPPER - October 1987 - Volume 9 Number 3 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 - October 1987 - Volume 9 Number 3 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 - October 1987 - Volume 9 Number 3 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 - October 1987 - Volume 9 Number 3 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 - October 1987 - Volume 9 Number 3 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 PAGESWAPPER - October 1987 - Volume 9 Number 3 VAX Systems SIG Fall 1987 SIR Ballot VAX Systems SIG Fall 1987 SIR Ballot DECUS membership number __________________ (six digits) Our site uses the following VAX cpus (check all that apply) 8700/8800 ___ 8600/8650 ____ 8500/8550 ____ 8300/8200 ____ 11/780,11/782,11/785 ____ 11/750 ____ 11/730,11/725 ____ MicroVAX I,II ____ MicroVAX 2000, VAXstation 2000 ____ We use VAXes in the following applications(Check all that apply) Business EDP ____ Software Development ____ Education ____ Computer Science Research ____ Data Acquisition/Control____ CAD/CAM ____ Service Bureau ____ Hardware Development ____ Scientific/Engineering ____ Office Automation ____ Telecommunications _____ Other __________________________ I support the following as the most important System Improvement Requests. (List from zero to fifteen SIR's): -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- I oppose the following SIR's as detrimental. (List from zero to five SIR's): -------- -------- -------- -------- -------- Mail to: Mark D. Oakley Battelle Columbus Division Room 11-6008 505 King Avenue Columbus, OH 43201-2693 USA To be counted, your ballot must be received by October 30. PAGESWAPPER - October 1987 - Volume 9 Number 3 VAX Systems SIG Fall 1987 SIR Ballot Tear out or photocopy reverse to vote on SIRs Mark D. Oakley Battelle Columbus Division Room 11-6008 505 King Avenue Columbus, Ohio 43201-2693 USA