.NONUMBER .LM 0 ^PY^- .PAGE SIZE 58,85 .LM 10 .RM 75 .NO FILL .NO JUSTIFY # .SKIP 5 .CENTER The RSX Multi-Tasker .CENTER April, 1987 .SKIP .CENTER ###^IS144^G"Fiat Hax"^IS204^G .SKIP .CENTER Fine Realtime Commentary Since 1975 .SKIP 6 .CENTER ^&Table of Contents\& .SKIP 2 .TAB STOPS 65 Food for Thought RSX-1 The Editor's Corner RSX-1 Running Boards RSX-2 Submitting Articles to the Multi-Tasker RSX-3 And That's The Way Things Are RSX-3 RSX SIG Election Nominations Open RSX-3 Response to Improving Spooler Performance RSX-4 The Notebooks of Justin L. Hewser RSX-6 Fall 1986 Q _& A Session Transcript RSX-8 Spring 1987 "Woods" Meeting Report RSX-24 .JUSTIFY .FILL .SKIP 15 .LM +5 .RM -5 Opinions expressed in the editorial section of the Multi-Tasker are those of the Editor. They do not represent the official position of the RSX SIG or that of DECUS leadership in general. .LM -7 .RM +7 .PAGE .COMMENT .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT Food for Thought .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT # .SKIP 7 .AUTOPARAGRAPH .CENTER ^&Food for Thought\& .SKIP "O civili! Si ergo. Fortibus es in ero." .BREAK "O nobili! Deis trux. Vatis enim, causan dux." .SKIP 2 .INDENT 30 ^IS144^G- Justin L. Hewser^IS204^G .INDENT 30 ^IS144^G##Latin for the Masses^IS204^G .COMMENT .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT The Editor's Corner .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT .SKIP 9 .CENTER ^&The Editor's Corner\& .SKIP .CENTER Bruce R. Mitchell .SKIP By the time you have this issue in your hands, the elections for the DECUS Board of Directors will be well underway. The entire RSX SIG Steering Committee encourages you to exercise this opportunity to influence the future of DECUS. Vote. Return those cards. As they say in Chicago: "Vote, and vote often." The SIG also has elections coming up. See the announcement by Ralph Stamerjohn in this issue regarding RSX SIG Executive Committee elections. If you are interested, run for a position yourself. Sure, it demands a lot of time, specifically Symposium and Woods meeting attendance. But this is your chance to shape the future of the RSX SIG. If you can't run, at least vote as you go by. The Spring Symposium is coming up pretty fast on us now, and, as always, the SIG has a splendid lineup of presymposium seminars and sessions. Some of the sessions include: .SKIP o##The Files-11 On-Disk Structure .BREAK o##The First National RSX Test .BREAK o##Real-Time Example Applications Panel and, of course the Software Clinic, the Q _& A session, "Meet the Developers", "What Digital Wants to Know About RSX", "System Programming Short Notes", and Magic aka "The Lore of RSX". Make ^&sure\& you get to Magic at Nashville. The SIG has a limited supply of ^IS144^G^IS204^G as prizes for the best presentations at Short Notes and Lore. These are similar to, but not the same as the bigger ^IS144^G^IS204^G that the SIG will be selling on a test basis. (Fair warning ahead of time: The big ^IS144^G^IS204^G are in very short supply - only 20 of them - so if you want one, make sure you get it ^&early\& in the week. Even if they go over big, there's no guarantee that we'll be able to get them into the DECUS store in Anaheim.) If you're still having trouble getting your supervisor to send you to DECUS, point out that the $500 registration is quite a bit less than the $1000-and-up that most 2 and 3 day training courses run. Shucks, take a week of vacation if you have to, sign up for a SuperSaver and come as you are. There is no better time and place to talk RSX with people who know it than at Symposia. The Editor is so happy he could just cry. He has received, over just the last week, more articles than he has seen over the last three months. Thanks to each and every respondent. Your copies of the limited edition prints will be in the mail long before this issue hits the streets. Keep them articles coming; there are still about 30 of the prints of the Fall 1986 SIG button artwork to give away. These prints are a limited edition, all signed by the artist and certified. The only way you can get one is to contribute an article. And, as always, if you don't want to ^&write\& an article, at least drop a line telling what kind of articles you'd like to ^&see\& in the Multi-Tasker. And now, it's editorial time. We're terribly sorry that Justin L. Hewser, our guest editorialist, couldn't get in to the offices this month to write an editorial. We changed the locks on all the doors and bought a mean guard dog. The Editor has been oiling and cleaning his terminal every 50,000 lines anticipating this changeback. So here comes this month's editorial, at Mach Two. .TEST PAGE 5 .SKIP 2 .CENTER ----- Running Boards ----- Of late, a number of "end users" have expressed misgivings about the directions taken by DECUS as an organization (see past editorials). While the Multi-Tasker is officially neutral on such political items, the Editor does have a few opinions on the topic. And - guess what? - he's going to share them with you. There is an election coming up for the DECUS Board of Directors. This is the group that sets the direction for DECUS as a whole. The people elected to these positions have the responsibility of representing and acting on the will of the general membership. If the people in those positions represent the membership, then those positions should be equally open to ^&all\& the membership. That seems logical, does it not? Well, this year an unnamed entity within the Society reputedly took upon itself the responsibility of pre-selecting a slate of candidates for the election. This certainly does prevent confusion among the electorate. But it also disallowed the candidacy of three well-known DECUS members who wish to run. These "dark horses" are not Board members now, so they were forced - on very short notice - to go to the membership at large with petitions to support their candidacy. Could you scrounge up a hundred DECUS members in your area in the space of a few weeks? Could you circulate a petition that fast, get it signed, and return it to DECUS? They tried. It seems unnecessary for us to force people who want to run for a DECUS position to go through that trouble. They are, after all, offering us a service. The Society should be grateful for their offer to serve. Even if the Society finds it necessary to restrict the number of candidates for elections, it seems equitable to demand that ^&all\& candidates be subject to the same rules. ^&You\&, gentle readers, are the membership of DECUS. When the ballots come out, it is ^&your\& votes that determine the outcome. Cast them for people who support the way ^&you\& want DECUS to be in the future. .TEST PAGE 5 .SKIP 2 .CENTER ----- Submitting Articles to the Multi-Tasker ----- Please submit machine readable media if you can. RX01, RX02, RX50, or 9 channel magtape at 800 or 1600 BPI are best. Any RSX volume format is acceptable except ROLLIN or PRESRV. ANSI, BRU and DOS FLX formats are well-liked by the Editor's tape drive. The Editor can now Kermit articles out of other hosts into the Multi-Tasker host. The reverse is unfortunately not true; the Multi-Tasker host is normally an isolate. If you want to submit an article via Kermit, call beforehand with (1) a login for the host machine and (2) system uptimes. Submissions which aren't machine readable take longer to get into print. The editor is lazy and types mass quantities only once a month when progress reports are due. If you preformat a submission in RUNOFF format, please set page size 58,80; left margin 10; right margin 75; and, when changing margins, use incremental changes rather than absolute. The editor blesses you for the consideration. Send all submissions to: .SKIP .NO FILL .NO JUSTIFY Bruce R. Mitchell Machine Intelligence and Industrial Magic PO Box 816 Byron, MN 55920 (507)#775-6268 .JUSTIFY .FILL .TEST PAGE 5 .SKIP 2 .CENTER ----- And That's The Way Things Are ----- _... this month in Pool Lowbegone, where the LL(1) grammars are strong, the NDFSM state tables are good-looking, and the recursive-descent stack depth is above average. .COMMENT .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT RSX SIG Election Nominations Open .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT .TEST PAGE 15 .SKIP 6 .CENTER ^&RSX SIG Election Nominations Open\& .SKIP .CENTER Ralph Stamerjohn .CENTER Meridian Technology Corp. .CENTER PO Box 2006 .CENTER St. Louis, MO###63011 .SKIP It is once again time for RSX SIG Executive Committee elections. This is official notice of the opening of nominations for elected Executive Committee membership. The RSX SIG policy-making group is the Executive Committee. It consists of 5 elected members and 2 appointed members. The SIG holds biannual elections to select a new group of elected members for the Executive Committee. There are two qualifications for these positions. The first is simple; you must return a petition with 5 or more RSX SIG members' signatures to the Nomination Committee. The second qualification is willingness and availability to serve. These positions demand significant commitment. Executive Committee members must attend both Symposia and both SIG Woods meetings each year. This is a direct commitment of 20 days. There is also an indirect commitment of about half an hour per day in phone calls and electronic mail - another 16 man-days a year. Employer approval and support are clearly required for this level of commitment. This is especially true because there is little compensation for these positions. Executive Committee members pay their own way to Symposia. Only the Woods meetings are covered by DECUS. If you are willing and able to serve in this capacity, send a simple petition with signatures and DECUS membership numbers of at least 5 SIG members to me, at the address above. And get them in soon, because nominations close two weeks after the Spring Symposium. .COMMENT .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT Response to Improving Spooler Performance .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT .TEST PAGE 15 .SKIP 6 .CENTER ^&Response to Improving Spooler Performance\& .SKIP .CENTER Fred Tromans .CENTER Central Electricity Generating Board .CENTER Bedminster Down .CENTER Bridgwater Road .CENTER Bristol BS13 8AN .CENTER England .SKIP ^IS144^GThe following article is a response to the article in the September 1986 issue about improving M-Plus spooler performance. Thanks to Fred for his work on this topic.#--##The Editor^IS204^G .SKIP The article in the September Multi-Tasker about improving M-Plus spooler performance was very helpful. The problem with completely flattening the spooler is soluble. You can build a totally flat spooler if you change the segment name in the GBLPATs. TKB takes the segment name for each overlay from the name of the first module in the overlay. When you flatten the despooler, segment INIT disappears, so the GBLPATs for segment INIT fail. However, if you change the GBLPATs to use segment RECEIV, it all works correctly again. Here is an example TKB command file for a fully flattened spooler. The changed GBLPATs are in italics. .SKIP 2 .NO FILL .NO JUSTIFY SY:[1,54]LPPFSL/PR/SL/-FP/CP/-IP/MM, MP:[1,34]LPPFSL/-SP = SY:[1,24]LPP/LB:RECEIV:POTS:INIT:JOBSTR:JOBEND:IOPRT:FILPRO:CONT SY:[1,24]LPP/LB:ERPRT:PRT:FLPAG:FLINI:JBINI:FLGEN:PRTDN:DONE SY:[1,24]QMG/LB:QMGSYM:FPRIV:PRCO SY:[1,54]RSX11M.STB/SS / TASK=LP0 IDENT=04.01 STACK=64 UNITS=6 ASG=CO:2 PAR=GEN:0:0 PRI=70 GBLDEF=E$OJFF:0 SUPLIB=FCSFSL:SV GBLDEF=$DELCK:0 GBLDEF=GLBUFZ:400 EXTSCT=INPBF1:402 EXTSCT=INPBF2:402 ^IS144^GGBLPAT=RECEIV:FRMTBL:144102 GBLPAT=RECEIV:FRMTBL+2:50102^IS204^G GBLDEF=Q$LSPS:174 GBLDEF=Q$HSPS:177 GBLDEF=Q$LSPF:3 GBLDEF=Q$HSPF:6 ^IS144^GGBLPAT=RECEIV:FRMTBL+6:102102 GBLPAT=RECEIV:FRMTBL+10:50102 GBLPAT=RECEIV:FRMTBL+12:102102 GBLPAT=RECEIV:FRMTBL+14:50102^IS204^G // .JUSTIFY .FILL .SKIP The flat spooler task image produced by this build procedure is significantly smaller. It also executes faster in a heavy spooling environment. Additionally, if you want to implement big-buffering as discussed in the original article, the flattened spooler still has sufficient free space available to implement it (although not as much as in the overlaid version). .COMMENT .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT The Notebooks of Justin L. Hewser .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT .TEST PAGE 15 .SKIP 6 .CENTER ^&The Notebooks of Justin L. Hewser\& .SKIP .CENTER Jim Bostwick .CENTER Cargill, Inc. .CENTER PO Box 9300 .CENTER Minneapolis, MN###55440 .SKIP This month continues our series of articles based on the personal programming notebooks of Justin L. Hewser. We all know Justin: he has been seen in nearly every programming shop. The desk in the corner, right next to Murphy. Yeah, him. Justin is a prolific programmer, and we have all at one time or another been forced to work with his product. Forced, because we'd sooner write an ACP in FOCAL than look at this stuff voluntarily! This month's note is from the MACRO section of the Notebooks. Justin apparently got hip to error detection somewhere along the line. This note describes his "generic" directive error handler. .SKIP 2 .LM +10 .RM -5 ^IS144^G*M330. -- MACRO Error Handler .SKIP It is a good idea to check for errors in system directives. This is done by looking at the directive status word, $DSW. The carry bit is also clear if the directive succeded. .NO FILL .SKIP XYZ$S _#EFN ; Some directive BCC 10$ MOV $DSW, R0 ; Make DSW show IOT ; Crash and burn 10$: .FILL .SKIP If an error is detected, the error code is put in R0, and the program dies. You can see the error code from the register dump.^IS204^G .LM -10 .RM +5 .SKIP The gig here is that the IOT (I/O Trap) instruction is illegal in user mode, and it causes the program to SST trap. The program probably wasn't processing SSTs, so TKTN then gets control, kills the task with the classic error message and dumps the general register set. Well, at least this is some form of error detection. In fact, I used to do something similar myself. I even took the trouble to put a tag in R1 which told me ^&where\& the program died without looking in the Taskbuilder map. _... which is one of the problems with this technique. You have to look in the map to see where the program died. If used in a utility subroutine, you still don't know where in the main program things went awry, just the utility that bombed. Another major problem with this technique is that RSX now uses the IOT for M-Plus V3.0 fast mapping. ^&Very\& strange things may happen if you fire an IOT in that environment! Aside from these considerations, the whole thing assumes that TKTN, the Task Terminator, is installed, TKTN is the guy who delivers those little messages giving the exit reason and register dump. No TKTN, no messages. Perhaps this whole idea came from the RSX Executive itself. RSX has a macro called "CRASH" which does an IOT. It is used whenever Executive code gets into a "Catch-22" situation, and can do nothing but roll over and die. Because RSX tries to detect as many errors as it can, the CRASH macro appears throughout the Executive sources. The astute reader of RSX, however, will note that CRASH is only used for truly irrecoverable errors - and that RSX does a lot more error detection and reporting in less final and more informative ways. There is really no excuse not to do reasonable error reporting. Justin was either just too lazy or believed his code wouldn't encounter errors. Come to think of it, Justin still ^&does\& believe his programs - and the systems they run on - are error free. We mortals know better, and should prepare for the inevitable. Ralph Stamerjohn has written an excellent, perhaps definitive, paper on error reporting. You will find it in the December, 1986 issue of the ^&DEC Professional\&. The associated code is also on the Fall 1986 RSX SIG tape. Bruce Mitchell also put his SST reporter/handler in a previous MACRO ^&Bag of Tricks\& column. It deals with non-directive errors, and is highly recommended. If you haven't written your own error handlers, use these. Better yet, incorporate them into SYSLIB - Ralph tells how - and dictate that ^&everyone\& (Justin included) use them. There is a wealth of information in Justin's notebooks. I have only begun to analyze them in detail. Here are some candidates for future columns: .SKIP o##Interactive Batch Processor .BREAK o##Easy TTDRV Patches .BREAK o##FORTRAN Decompiler .COMMENT .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT Fall 1986 Q _& A Session Transcript .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT .TEST PAGE 15 .SKIP 6 .CENTER ^&Fall 1986 Q _& A Session Transcript\& .SKIP .CENTER Gary Maxwell, editor .CENTER U.S.#Geological Survey .CENTER Menlo Park, CA .SKIP The RSX and P/OS Technical Question and Answer session occupied three full hours on Tuesday afternoon during the Symposium. In contrast with previous versions of this session, no Q/A forms were used. This resulted in a somewhat informal and spontaneous session, with many technical tidbits flying about. This transcript is taken from the audio recording of the session. Text enclosed in square brackets contains audience comments and follow-up information from the editor. .left margin +3 .SKIP 2 .INDENT -3 Q##Under RSX, how can I set one of the terminal ports to be 7-bit and even parity from a program? .SKIP .INDENT -3 A##This can be done using the Set Multiple Characteristics QIO directive. Consult the Full Duplex Terminal Driver section of the I/O Drivers Reference manual. [A session was given later on terminal driver functions.] .SKIP 2 .INDENT -3 Q##Under M-Plus V2.1, using BRU with verify on a TK25, we encountered multiple "Device Not Ready" and "Fatal Hardware" errors during the verify pass. Is this a known problem? It happens on all of our TK25s. .SKIP .INDENT -3 A##We don't know. It sounds like a hardware problem, since both of these errors are generated by the device driver. Check your hardware configuration. .SKIP 2 .INDENT -3 Q##How do you make a bootable BRUSYS tape for the TK50? I have a BRUSYS tape at 800 BPI, can I just use TPC to copy it over? Can I use VMR to create a BRUSYS tape under VMS? .SKIP .INDENT -3 A##You should be able to copy the BRUSYS tape onto the TK50 and be able to boot it. We don't think you can VMR SAV a BRUSYS tape on the VAX (but we don't know why not). .SKIP 2 .INDENT -3 Q##Do you have any plans for a Pascal-11 compiler with RMS support under RSX? .SKIP .INDENT -3 A##RMS is one of the highest priority items with the Pascal developers. .SKIP 2 .INDENT -3 Q##I need a DEUNA driver to run on a cramped 11/45, and I am running RSX V3.1. Is it possible to write one? I do not have enough available memory for DECnet. .SKIP .INDENT -3 A##You can write your own driver; however, the DEUNA is a complex controller. Consider upgrading to V4.2 (leaving out unnecessary options during SYSGEN), and use the DEUNA driver (XE:) supplied with the RSX kit. .SKIP 2 .INDENT -3 Q##When using third party MSCP controllers and drives, attempting to boot unit zero with unit one spinning up causes SAV to hang after the first message, until unit one is ready. The controller vendor claims it is an RSX problem. Is it? .SKIP .INDENT -3 A##When bringing the device online, SAV may be intentionally waiting for the second unit to spin up, since SAV makes no assumptions about the drive type (it may be an RC25). Your controller may also behave differently than a DEC controller while a drive is spinning up. .SKIP 2 .INDENT -3 Q##When FAL is invoked over the network to process large directories, secondary POOL becomes depleted. It remains that way until a later, simpler operation is performed, and secondary POOL recovers. What is happening? .SKIP .INDENT -3 A##[Every time FAL accesses a file, it sends a message to FALLOG. FAL doesn't call the logger (FALLOG) until it exits. If FAL aborts, FALLOG may not be called. You can solve the depletion problem when it occurs by issuing a RUN FALLOG command. The general workaround is to disable FAL logging by rebuilding FAL after setting the symbol $DFLLG to false in [137,10]DECPRM.CMD on the DECnet distribution. This problem should be fixed in Update C or D.] .SKIP 2 .INDENT -3 Q##I am running into POOL depletion problems on 11M when I try to maintain a large System Task Directory and run DECnet at the same time. Removing applications tasks from the STD causes problems because the applications have to run from privileged accounts for non-privileged users. Are there any workarounds? .SKIP .INDENT -3 A##Besides recommending a system upgrade, we suggest that you modify the dispatcher (MCRDIS) to perform special processing of non-privileged UICs to obtain a workaround. .SKIP 2 .INDENT -3 Q##What changes can we expect to P/OS and the Tool Kit between V3.0 and V3.1? .SKIP .INDENT -3 A##V3.1 is an update of V3.0. It does not require reinstallation of the system. Major changes primarily involved correcting print server bugs. There is also an update to the Tool Kit. .SKIP 2 .INDENT -3 Q##We recently upgraded from M to M-Plus. We assigned two LUNs to the terminal, 5 and 6, from a heavily overlaid task. Attempting to read from LUN 6 caused a system crash. Flattening the task or using only one LUN solved the problem. .SKIP .INDENT -3 A##Please send a crash dump along with an SPR. We don't know what is happening. .SKIP 2 .INDENT -3 Q##We upgraded a 10 Megabyte drive to a 30 Megabyte drive. The new drive does not work. Why not? .SKIP .INDENT -3 A##The drive's select jumpers are probably incorrect. If this is in a BA23 Micro box, the first Winchester must be set to unit 3. .SKIP 2 .INDENT -3 Q##We are running a medical Accounts Receivable package under A-to-Z on Micro/RSX. On occasion, the console terminal and the application hang. When we log onto another terminal, and do a SHO TERM command for the console, we get the message "Device not a terminal." What is the problem? .SKIP .INDENT -3 A##It sounds like the console is being spooled. Check for an active task called TT0, which could be a spooler task. .SKIP 2 .INDENT -3 Q##I recently copied and initialized a system on a new RC25 drive. Following the virgin boot, I tried to SAV the system. SAV complained that RCT had active I/O. When I retried the operation about 10 seconds later, it worked. Why? .SKIP .INDENT -3 A##When the driver first initializes itself (after the virgin boot), the driver invokes RCT to read the factory replacement control table on the drive. This can take some time. Just keep typing SAV until it works. .SKIP 2 .INDENT -3 Q##Under M-Plus V3.0, continually hitting the down arrow key depletes secondary POOL and eventually crashes the system. Is this a known problem? I brought a crash dump. .SKIP .INDENT -3 A##The system should not crash, but the performance will degrade. We would like to see the crash dump. .SKIP 2 .INDENT -3 Q##After performing a SET HOST from a VMS system to an RSX system, terminal output wraps at column 72. Displaying the RT: device buffer shows 255 characters, with NOWRAP. What's the problem? .SKIP .INDENT -3 A##VMS and RSX do not correctly pass all terminal characteristics to each other. Buffer size is probably one of these. We are still working on this problem. One workaround is to set the buffer size before the SET HOST command. .SKIP 2 .INDENT -3 Q##In trying to use PRO/LAT, I cannot stay connected for more than one or two minutes at a time to an RSX node before I get disconnected. It works fine with VMS, and a server on the same cable works fine to both hosts. .SKIP .INDENT -3 A##The latest update to DECnet seems to improve the situation. .SKIP 2 .INDENT -3 Q##Has the positional device driver which used to be packaged with P/OS been upgraded? .SKIP .INDENT -3 A##It didn't make it on the V3.0 kit. It will be out with the next version of PRO/Sight, with the next version of P/OS. .SKIP 2 .INDENT -3 Q##Under V4.1 of RSX, we installed the Shuffler in the GEN partition and made it checkpointable. The system crashed after about 15 minutes with a kernel stack underflow. There appears to be code in the Shuffler to prevent this, but it doesn't seem to be working. Has this been fixed for V4.2? .SKIP .INDENT -3 A##Probably not. .SKIP 2 .INDENT -3 Q##There is a bug in EDT on M-Plus which causes it to crash in the system-wide initialization file. The only workaround that seems to work is to build EDT without any of the M-Plus features (/MU, etc.). When will this problem be fixed? .SKIP .INDENT -3 A##We don't know. Not a lot of development effort is being spent on PDP-11 EDT. There are changes to EDT for Update C of V3.0. Try the PDP-11 Languages and Tools Q_&A session later in the week. .SKIP 2 .INDENT -3 Q##With the PRO Host Tool Kit going away in the near future, is it feasible to build PRO applications under the vanilla M-Plus system? Will things on the Tool Kit be supplied on M-Plus that aren't on the Tool Kit? What about FDT? .SKIP .INDENT -3 A##Everything you need to build an application can be done with the M-Plus Utilities, as long as they point to the P/OS-specific libraries. PRO/F77 has different object libraries and resident libraries. FDT has not been changed or rebuilt in some time; it should continue to work. We don't have all the answers as to the problems that will be caused by the retirement of the Pro Host Tool Kit. .SKIP 2 .INDENT -3 Q##We just received two RA60's, and we would like to do a SYSGEN on them. We have an RL02-based distribution. Can we just change a few symbols to make SYSGEN think it's a big-disk distribution? It seemed to work when we tried it. .SKIP .INDENT -3 A##That trick should work. If it wasn't done right, then SYSGEN would fail early in the procedure. .SKIP 2 .INDENT -3 Q##On P/OS V3.1, Pro LOAD tries to bring all units online when it loads a driver. My driver does not want to come online until a task establishes a link with it. When my driver rejects the online request, Pro LOAD prints an error message and unloads the driver, which is not what I want. Can this behaviour be changed? .SKIP .INDENT -3 A##Try one of the following: modify your application to work around the way Pro LOAD works; have the driver remain offline while lying to Pro LOAD by returning success; or during the online call, have the driver shuffle the kernel stack so that it gets called back to turn itself offline after Pro LOAD finishes the online call. .SKIP 2 .INDENT -3 Q##I can perform a standalone SYSGEN with my distribution on a Cipher M891 (MS:) tape drive, but when I boot the new system, I can't access the tape drive anymore. During SYSGEN, the drive appears as a TK25; after booting the new system, it shows up as a TSV05. What's wrong? .SKIP .INDENT -3 A##There was a problem with MSDRV for I/D space systems on 22-bit Q-Bus systems. It was fixed in V3.0 Update B. .SKIP 2 .INDENT -3 Q##We have users queueing up a whole bunch of print jobs while using A-to-Z, and they regularly call us to delete print jobs from the queue. Is there an easy way of deleting all jobs for a particular user? .SKIP .INDENT -3 A##Try creating a global command file which a user can invoke which will walk him/her through the process of deleting queue entries. You might also talk to the A-to-Z developers about adding a menu option to do this. .SKIP 2 .INDENT -3 Q##We currently have devices on DL11-Ws, where we switch between a custom device driver and the terminal driver using a set of commands. If we change from a DL-11 to a multiplexor, can we switch a few of the ports on the multiplexor to the special driver, or do we have to switch them all? .SKIP .INDENT -3 A##In general, you have to switch all of the ports. You can try to create a mechanism whereby your driver fielded all interrupts, and then simulated an interrupt to the terminal driver, when necessary. That is a non-trivial solution. .SKIP 2 .INDENT -3 Q##The Connect to Interrupt and Switch State Directives do not seem to work for an I/D space task. A mapping error occurs when APR 5 has I- and D-space segments. .SKIP .INDENT -3 A##This is intentional. One way to enable I- and D-space when your system state code is entered is to read the User Mode D-space APRs to load up the Kernel Mode D-Space APRs. The Executive really doesn't care about the APR contents once your code gets control. (Try to leave APRs 0 and 7 alone, though.) .SKIP 2 .INDENT -3 Q##I was able to build loadable XDT from the available tools on the M-Plus distribution, and it worked. Will you support it as a SYSGEN question? What about loadable XDT on 11M? What about loadable crash dump driver support? .SKIP .INDENT -3 A##The loadable XDT source code has conditional code in it for 11M, but it hasn't been tested. On 11M, loadable XDT is run as a task (and not loaded) and it "disappears" into the system. Loadable crash dump driver support would take some work to implement on M-Plus. Loadable XDT does not have to be a SYSGEN option: XDT is a vectored task image, so you can always load it, and it is transportable to any other system. .SKIP 2 .INDENT -3 Q##We have some applications using the terminal driver for non-terminal devices, and recently the controllers hung. Examination of the UCB showed a bit called "terminal disabled" was on. What is it and what does it mean? .SKIP .INDENT -3 A##The disable line is used for /REMOTE terminals and LAT terminals to perform modem control. It should not be set unless the terminal was set /REMOTE. .SKIP 2 .INDENT -3 Q##What are the new features of P/OS V3.1? Does it affect software compatibility from V3.0? .SKIP .INDENT -3 A##Changes were made to Print Services, there were some Base System fixes, and some Tool Kit changes were made. There shouldn't be any incompatibilities. .SKIP 2 .INDENT -3 Q##I have a user who was able to use CTRL/C to abort his current program on a Micro/RSX system. How do I turn it on for M-Plus? .SKIP .INDENT -3 A##It only works for DCL. You need to specify the /CTRLC switch when DCL is initialized with the CLI command, and then set each individual terminal to /CTRLC. .SKIP 2 .INDENT -3 Q##I have been trying to "cluster" FCSRES against an I/O Page Common. I do it by having TKB map FCSRES to APR 6 and IOCOM to APR 7. The the program performs a Get Mapping Context, and then maps FCSRES and IOCOM when each is needed. I get an alignment error the first time I try to remap, but the operation works anyway (crash dump verifies this). Do you know if there are any problems with doing this? .SKIP .INDENT -3 A##Try to create a simple case, and send in an SPR. It should work. .SKIP 2 .INDENT -3 Q##Is there any difference between tasks which link to FCSRES with ANSI magtape support and FCSRES without ANSI magtape support? Are tasks which link to RMSRES transportable between 11M and M-Plus, since the number of RMS resident library segments is different between the systems? .SKIP .INDENT -3 A##For FCSRES, the task images will be the same, since the vectors for both flavors of FCSRES are identical. For RMSRES, we're not sure. .SKIP 2 .INDENT -3 Q##Our system degrades after a few months of use, apparently due to file fragmentation on the disk as our files grow. Besides a backup and restore, how can we reorganize a disk volume? .SKIP .INDENT -3 A##We have seen and tried several schemes for online compression, some better than others. There is a commercial second-source product out, but we personally wouldn't try online compression on a disk that we cared about. .SKIP 2 .INDENT -3 Q##Do Print Services under P/OS V3.1 allow RS/1 to print? It doesn't work under V3.0. .SKIP .INDENT -3 A##Check the Release Notes for V3.0 to see if there was a restriction. We will try to figure out if there is a problem. .SKIP 2 .INDENT -3 Q##A user on my system submitted a trivial batch job. A few hours later, the batch job was still active, the Virtual Terminal was logged in, but no tasks were active. The log file showed a message that error status was returned for something and the batch job was being terminated, but it didn't finish. What happened? .SKIP .INDENT -3 A##It looks like BYE couldn't get started up by the Batch Processor. We're not sure what really happened. .SKIP 2 .INDENT -3 Q##On an 11/84, we had problems running BAD on an RA81. BAD would run for about 15 minutes, then it would "die horribly" [whatever that means], and the RA81 would be left software write-protected. I finally got it to work by submitting and resubmitting an overnight batch job to run BAD on the disk, and now everything works fine. Any ideas? .SKIP .INDENT -3 A##There was a problem with the Replacement Control Task for V3.0 Update B, one effect being the software write-protecting of the disk. It has been fixed for Update C. .SKIP 2 .INDENT -3 Q##We are upgrading from 11M V3.1 with Fortran-IV-Plus to M-Plus and Fortran-77, and have been told about some differences between the Fortran/Macro interface, especially with logical variables. What should we be concerned about? .SKIP .INDENT -3 A##An Appendix to the Fortran-77 manual has a list of differences between F77 and F4P. There are some incompatibilities. You can also use the /NOF77 switch to F77 to turn off some of the language features. .SKIP 2 .INDENT -3 Q##I couldn't install V3.0 of P/OS on a 5 Megabyte Pro 350, because I ran out of disk space. Can it be done? .SKIP .INDENT -3 A##We have a big disk space crunch on V3.0. Come to a session tomorrow which will cover how to optimize the disk space. You should be able to get the base system, Pro Comm, and DCL on the disk, but not all that and the Tool Kit. .SKIP 2 .INDENT -3 Q##How many blocks are needed for a single dedicated application running on P/OS? .SKIP .INDENT -3 A##The documentation describes how to build a dedicated application that runs off the floppies. P/OS needs a minimum of about 450 blocks to boot the system, and up to 1000 blocks are generally needed for overhead. It's up to you to do the rest of the tailoring. .SKIP 2 .INDENT -3 Q##We perform a SET HOST to a Micro/RSX node from a VAX node. After we log in, a Control-C hangs the terminal port. TSC reported that it is fixed in V3.1 of Micro/RSX, but apparently it isn't. Do you have anything new on that? .SKIP .INDENT -3 A##We don't know if it has been fixed. .SKIP 2 .INDENT -3 Q##Can V3.0 of M-Plus be generated on an online V2.1 system? .SKIP .INDENT -3 A##Yes. .SKIP 2 .INDENT -3 Q##What does the block mode attribute of the TT driver do? .SKIP .INDENT -3 A##Absolutely nothing. .SKIP 2 .INDENT -3 Q##I also seem to have a problem with the Replacement Control Task software write-locking my disks. Data caching errors occur on the write-back. Is this the same problem? .SKIP .INDENT -3 A##It probably is. Until Update C arrives, try removing RCT and see what happens. You should get a hard error in that case. .SKIP 2 .INDENT -3 Q##When starting up Pro Comm, it takes over 60 seconds for the window manager to come up with the first menu. Is there any way to speed it up? .SKIP .INDENT -3 A##Disable the graphics option. See the Pro Comm documentation on how to do it from the setup screen. You can also call up DTE directly from the Tool Kit. .SKIP 2 .INDENT -3 Q##When you perform a Set Terminal /Inquire in Pro Comm, the cursor is positioned to the second column in the first row, and the disk makes noises. What is happening? .SKIP .INDENT -3 A##You don't want to know. [Laughter] The system goes to the disk for the Regis initialization codes. .SKIP 2 .INDENT -3 Q##If you disable graphics on the Pro, do you lose VT125 emulation? .SKIP .INDENT -3 A##Yes. You lose all Regis emulation. .SKIP 2 .INDENT -3 Q##Does Pro Comm V1.7 or V1.8 work on P/OS V3.0? .SKIP .INDENT -3 A##Absolutely not. .SKIP 2 .INDENT -3 Q##We would like to get to the video registers on the Pro. Do you have documentation on the terminal subsystem so that we can do our own graphics? .SKIP .INDENT -3 A##We originally distributed the documentation, and we took it away. You have to be extremely careful so that the system doesn't crash. See us offline for some information. A real neat thing you can do is to run tasks in the bitmap partition. [Laughter] .SKIP 2 .INDENT -3 Q##When M-Plus V2.1C came along, we rebuilt all of our tasks to take advantage of vectored FCSFSL, and thought we'd never have to build them again. Along came V3.0, and we want to avoid rebuilding everything again. What can we do? .SKIP .INDENT -3 A##The news is that V2.1C FCSFSL really wasn't vectored. We thought it was, and advertised it as such. You can still install the old one. We think we've got it right this time, but you'll have to rebuild your tasks for the last time. .SKIP 2 .INDENT -3 Q##When I upgraded a system from 11M V4.1C to M-Plus V3.0 on an I/D system, the Basic-Plus-2 compiler could no longer open its work files. When I installed the compiler with a POOL-resident header, it worked. Do you have an explanation? .SKIP .INDENT -3 A##We have no idea. Until you apply Update C, avoid installing tasks with POOL-resident headers, because issuing PLAS directives from such tasks will crash the system. .SKIP 2 .INDENT -3 Q##We want to boot 11M V4.2 off of an RX50. How do we do it? .SKIP .INDENT -3 A##It has to be a very small system. Load the drivers, then delete the driver images. Then install the privileged tasks that you need, fix them, and delete their images (if they are not overlaid). .SKIP 2 .INDENT -3 Q##Can P/OS support Anciliary Control Drivers? .SKIP .INDENT -3 A##No. .SKIP 2 .INDENT -3 Q##With Pro Comm V3.0 and Synergy, it would be nice to go directly into the menus. Can we get it like before? .SKIP .INDENT -3 A##No. .SKIP 2 .INDENT -3 Q##Are you aware of the problem with the new Pro Comm where you connect to a modem and the characters come in but are not displayed on the screen until you do a Clear Comm? .SKIP .INDENT -3 A##We aren't aware of the problem. .SKIP 2 .INDENT -3 Q##How often do you expect P/OS to be updated now? .SKIP .INDENT -3 A##We are currently working on V3.2. .SKIP 2 .INDENT -3 Q##The DCL SHOW ERROR /RECENT and /HISTORY options don't seem to be documented, but there are help files. They don't seem to work. What about them? .SKIP .INDENT -3 A##They are there for Micro/RSX, and they don't function under M-Plus. You can go into the M-Plus sources (SYSCM.MAC) and modify $EFMSK to change error logging behaviour to include more packet types. .SKIP 2 .INDENT -3 Q##Can you distribute two versions of P/OS: one for hardware that supports I/D space, and one for hardware that doesn't? .SKIP .INDENT -3 A##No. We don't have the development team to maintain and support two different versions of the system (four if you count Pro Server). .SKIP 2 .INDENT -3 Q##Regarding SET HOST problems between VMS and RSX, I cannot get 132 column mode to work from an RSX node on a VAX host. Can you get the RSX/DECnet group to talk to the VMS/DECnet group to solve these problems? .SKIP .INDENT -3 A##Noted. .SKIP 2 .INDENT -3 Q##We have the Ivis touch screen monitor on our Pro 350. When we start the verification application for the touch screen after the system has been booted, the application hangs. If we abort and restart the application, it works. What is the problem? .SKIP .INDENT -3 A##We don't know about the problem, but we'll look into it. .SKIP 2 .INDENT -3 Q##I am trying to get escape sequences into a CLI that I wrote, but I'm losing the characters. How can I get it to work? .SKIP .INDENT -3 A##Unfortunately, there is no attribute you can set to enable unsolicited input to have escape sequences. One workaround is to set up a CLI that does very little except to communicate with a task which keeps outstanding read requests with escape sequences enabled to the terminal. .SKIP 2 .INDENT -3 Q##We're having problems with the LAT driver. Terminal characteristics are not being properly reset (/ECHO, for example) when users disconnect. Attempts to use SYSLOGOUT.CMD to reset the terminal are only partially acceptable. Can this be improved? .SKIP .INDENT -3 A##Look in the terminal driver module TTMOD.MAC to add the ability to reset any terminal characteristics that require it. We will consider adding some more characteristics to the reset function. .SKIP 2 .INDENT -3 Q##I keep getting what I consider to be undeserved parity errors on my RQDX1/RD52 subsystem under Micro/RSX V3.0, but they aren't being logged by the error logger. What is the problem? .SKIP .INDENT -3 A##We don't know why you are getting so many errors. [Someone from the audience suggested that a peculiarity with the controller prevents the error logger from writing packets to disk while a retry operation is in progress.] .SKIP 2 .INDENT -3 Q##RMS always uses a file extension of one block [for sequential files], no matter what the volume default is. Why is this and is there a remedy? .SKIP .INDENT -3 A##RMS has a complicated algorithm for calculating the extent size. It has to consider file organization, bucket size, and other criteria. We have tried various changes to the algorithm; the current one is the best one we have. It would make sense for sequential files to use the default; we can't remember why that isn't so. .SKIP 2 .INDENT -3 Q##I recently tried to get help for P/OS from Telephone Support Center in Atlanta, and they had incorrect information for my problem. Is there any ongoing effort to enlighten these people? .SKIP .INDENT -3 A##Yes. We're not always successful. We suggest that if you don't get satisfaction from TSC, call them back and ask to have your problem elevated. The problem then comes to us, so that we can pass along the right answer. .SKIP 2 .INDENT -3 Q##A driver that we have for an array processor requires tasks to have POOL-resident headers. Could you expound on external headers and speculate on the restriction we have? .SKIP .INDENT -3 A##Until M-Plus V3.0 Update C, if there is any task which uses PLAS directives (including FCSRES, memory-resident overlays, and Fortran programs with Virtual Arrays) with a POOL-resident header, the system does not correctly initialize a location, which causes writes to random memory locations. Regarding your driver, it probably is missing the straightforward code used to map to an external header, which was a design decision on the vendor's part. .SKIP 2 .INDENT -3 Q##The documentation indicates that you can pass messages to a CLI task. Is there an Executive directive available for a task to send a message to a CLI? .SKIP .INDENT -3 A##We decided not to implement one at this time. We can do it if there is sufficient interest. The message facility in its current form is not very useful. .SKIP 2 .INDENT -3 Q##Is there some facility to allow a user to abort a remote terminal session over DECnet from an RSX host? .SKIP .INDENT -3 A##Yes. Type CTRL-_\ (Control-Backslash) and RETURN. You will get a local node prompt, and then type CLEAR HOST to break the connection. .SKIP 2 .INDENT -3 Q##When logging off remote terminals, we have encountered system problems when modem noise enters the system after BYE completes. Is there some way to remedy this? Slaving the terminal seems to be a possibility. .SKIP .INDENT -3 A##We suggest that you create a privileged task which spawns BYE and then resets the slave bit after BYE exits. .SKIP 2 .INDENT -3 Q##Regarding RMS file extension on the Pro, we have been field testing Pascal. When Pascal extends its work files, it extends them by ten blocks. However, if the largest fragment remaining on the disk is less than ten blocks, Pascal quits. FCS is able to give back what it can; why can't RMS? .SKIP .INDENT -3 A##The Pro version of RMS does extend files by a default value of ten blocks. We can't tell if it is Pascal or RMS which is causing this problem. If Pascal is asking for ten contiguous blocks, then RMS will return with an error if it cannot allocate that. If Pascal asks for the default extension, then RMS will return any ten available blocks it can find. We suggest that you talk to one of the languages people. .SKIP 2 .INDENT -3 Q##Regarding software licensing, if I use the coprocessor products (KXT11, KXJ11), I can end up with sixteen PDP-11's running RSX on one "system." What are my licensing options? .SKIP .INDENT -3 A##At this point, you need to purchase a Class L license for whichever system you want to run on the board or chip set you are using. We are looking into licensing options for these particular products, including a new variant of the old General License. We are open for input on this. [OK, folks, you heard it right. Send those letters to: Dick Day, Product Manager, Digital Equipment Corporation, 110 Spitbrook Road, ZK01-3/G10, Nashua, NH 03062-2642.] .SKIP 2 .INDENT -3 Q##In the file header format on RSX, there is a much-needed attribute called "Printable File." Have you considered this? .SKIP .INDENT -3 A##No. It would be difficult for the system to tell whether a particular file is printable or not. We leave it up to the user to decide that. .SKIP 2 .INDENT -3 Q##There are serious deficiencies with remote terminal support from VAX nodes to RSX nodes, including /INQUIRE support, prompting, extra line feeds, etc. What is the prognosis? .SKIP .INDENT -3 A##We have tried to clean up some of the problems. The real problem is a lack of communication between VAX-DECnet and RSX-DECnet people to get the VAX people to fix the problems. We suggest you send in an SPR to the VAX people. .SKIP 2 .INDENT -3 Q##Is there any documentation describing the various error messages that are reported by various devices; in particular, the DU driver? .SKIP .INDENT -3 A##The messages have a one-to-one correspondence with the errors described in the hardware error manual. The exception is the DU driver, which uses the proprietary MSCP protocol. .SKIP 2 .INDENT -3 Q##We have had a similar problem on Micro/RSX with RPT reporting unknown errors for the DU driver and "Device Message" packets that are indecipherable. It took us a long time to convince Field Service that there was a hardware problem. Can there be any improvement? .SKIP .INDENT -3 A##Even though you only have a DU type disk on your system, it looks like RPT get a packet it couldn't recognize. The Micro/RSX error library is heavily pruned, so it can't translate every error for every device. We're sorry, we can't give more information for the DU driver packets. .SKIP 2 .INDENT -3 Q##We are having problems with time-blocked batch jobs never getting started when the time comes due. The problem occurs randomly. Any ideas? .SKIP .INDENT -3 A##We don't know what the problem might be. [Audience suggestion: You can rebuild Indirect with Foreign Command Support by enabling the DEFCMD symbol. Install it, schedule it to run at a certain time, and then it looks up INDINI???.CMD where ??? are the first three characters of the task name.] .SKIP 2 .INDENT -3 Q##I would like to have a read-wait-and-lock feature for RMS. It would make applications requiring several communicating tasks easier to implement, so that a task can post a read on a shared file and wait for it to become unlocked. .SKIP .INDENT -3 A##We would have to put code in the Executive to support waiting for a particular block of a file to get unlocked: a difficult proposition. [Your best bet would probably be to use Group Global or Global event flags to have the tasks signal when a block is unlocked, so the other one can retry the read.] .SKIP 2 .INDENT -3 Q##Is there any way in RMS to tell it to update a record but leave the record locked? .SKIP .INDENT -3 A##We don't think so. That is all done as a part of reads and writes. .SKIP 2 .INDENT -3 Q##I've hooked up two applications terminal ports on a MicroVAX II to TT0: and TT1: on an 11/24. I then performed a SET HOST/DTE to TT0: and started up the print queue to TT0: on the 11/24. Things printed on the MicroVAX for awhile, and then it hung. I then performed another SET HOST/DTE to TT1: on the 11/24, hit a RETURN, and things started printing again for awhile. I continued to intervene until the print queue was empty. Trying the same procedure with Kermit worked perfectly. Any ideas? .SKIP .INDENT -3 A##There may be a problem at the VAX side, but we're not sure. .SKIP 2 .INDENT -3 Q##Is there any way to perform a BRU /IMAGE:SAVE/NOINI to a Files-11 mounted disk? The command refuses to be parsed correctly. .SKIP .INDENT -3 A##That may be a recent problem, because it used to work. .SKIP 2 .INDENT -3 Q##The DTEs for M-Plus V3.0A and for 11M V4.1D do not support wildcard filenames. Can this be fixed? .SKIP .INDENT -3 A##Update C for M-Plus should fix the problem. [No mention of 11M.] .SKIP 2 .INDENT -3 Q##How can you perform an image copy of an RX50 to another RX50? Could I use PRESERV? .SKIP .INDENT -3 A##[From the audience: use TPC, a DECUS utility]. PRESERV wouldn't know how to deal with RX50s. You could also write a block-for-block copy task, or use one of the Update tools on XXDP. DECUS may have a disk copy utility. .SKIP 2 .INDENT -3 Q##When trying to build a I/D multi-user task, the EXTTSK option to TKB seems to extend the read-write I-space, and not the read-write D-space. .SKIP .INDENT -3 A##That sounds like a problem. Please send in an SPR. [It seems to work on Update C.] .SKIP 2 .INDENT -3 Q##Do you give anything up when XDT is included in a system with Kernel D-space support, such as POOL? .SKIP .INDENT -3 A##About the only thing you give up is physical memory. XDT mostly uses I-space. Loadable XDT will be on the next version of M-Plus. .SKIP 2 .INDENT -3 Q##On 11S V4.2D, we would like to be able to install and fix tasks beyond 124K of memory. Any possibility of this happening? .SKIP .INDENT -3 A##No. We looked into it, and decided not to do it for various reasons. We can look into it again. .SKIP 2 .INDENT -3 Q##Regarding the Test Feature Directive, I am having problems determining if the task has Fast Map support. The same holds true for P/OS. .SKIP .INDENT -3 A##For M-Plus, the problem is fixed in Update C. For P/OS, we will try to fix the problem soon. .SKIP 2 .INDENT -3 Q##We would like to be able to use DTE from Batch to get into a terminal server to change the characteristics of the server. The problem is that DTE does not want to accept data from the batch stream. Why not? .SKIP .INDENT -3 A##DTE uses QIO functions that the Virtual Terminal driver does not understand. Your best alternative is to write a program that talks directly to the server. .SKIP 2 .INDENT -3 Q##When can we expect to get terminal server support for 11M? .SKIP .INDENT -3 A##We don't have a plan at present to do it. It is more technically complicated than we originally thought. .SKIP 2 .INDENT -3 Q##I'm getting lots of Data Access Protocol errors when performing wildcard copy operations between a MicroVAX II and an M-Plus system. Has anyone else seen this? .SKIP .INDENT -3 A##[Some others in the audience had seen the problem.] .SKIP 2 .INDENT -3 Q##Would it be possible to put the files necessary to create a BRUSYS tape on the M-Plus kit? I can't get BRUSYS to boot from an MM: device on an 11/44. .SKIP .INDENT -3 A##You would need sources for BRU, FMT, CNF, and BAD, plus an 11S system. That is just too much for us to put on the distribution. Regarding your problem, try lowering the baud rate of the console and check the XON/XOFF settings. .RM -3 .COMMENT .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT Spring 1987 Woods Meeting Report .COMMENT +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .COMMENT .TEST PAGE 15 .SKIP 6 .CENTER ^&Spring 1987 "Woods" Meeting Report\& .SKIP .CENTER Bruce R. Mitchell, editor .CENTER Machine Intelligence and Industrial Magic .CENTER PO Box 816 .CENTER Byron, MN###55920 .SKIP The 1987 RSX SIG Executive Committee planning meeting was held in Salt Lake City the weekend of January 31. These meetings (the "Woods" meetings) are held by the SIG leadership to determine the future direction of the SIG, to discuss current problems, and to bring members up to date on DECUS affairs. The following DECUS members attended this meeting: .SKIP Al Bennett - Exec Committee, Communications Committee rep .BREAK Jim Bostwick - Steering Comm., Working Group coordinator .BREAK Ed Cetron - Exec Committee, SIG Lobbyist .BREAK Bob Curley - Guest, DECUS Board representative .BREAK Dan Eisner - Exec Committee, RSX SIG chair .BREAK Beverly Kasper - Steering Committee, Suite Coordinator .BREAK Brian McCarthy - Guest, Digital RSX implementation group .BREAK Gary Maxwell - Exec Committee, Budget and Finance .BREAK Bruce Mitchell - Exec Committee, Multi-Tasker Editor .BREAK Rick Sharpe - Steering Committee, Symposium Committee rep .BREAK Bob Uleski - Exec Committee, Planning Coordinator .BREAK Denny Walthers - Exec Committee, PSS Coordinator .SKIP 2 .CENTER Saturday, January 31 The meeting began with a review of the agenda by Dan Eisner. Individual reports followed from the representatives to Communications Committee, Symposium Committee, Planning Committee and Finance Committee; and from the Multi-Tasker, Lobbyist and PSS Coordinator. Al Bennett (Communications) is now on the Budget subcommittee of Communications Committee. He pointed out that removal of the SIG notes purchase option from the pre-registration packet was done without our permission. Discussion followed on the usefulness and appropriateness of refereed papers at symposia. Denny Walthers (PSS) noted that one RSX PSS in Nashville will have a computer online for demonstrations. This is a first for the SIG. Gary Maxwell (Finance) submitted the FY88 activity plan for the SIG. The SIG is requesting $31K for FY88, up $6K from this year's $25K. Discussion of the budget followed. It was pointed out that the proposed standards activity will probably be removed. The prototype budget was adopted by unanimous consent. Gary's budget report was followed by a discussion of the upcoming RSX SIG elections. As 1987 is an odd-numbered year, the new SIG bylaws require an Executive Committee election this year. Dan Eisner reported that Ralph Stamerjohn agreed to accept the job of nomination committee chair for the upcoming SIG Executive Board elections. Bylaws require publication of a notice of open nominations in the Multi-Tasker; Bruce Mitchell will coordinate with Ralph. Dan Eisner stated that he is standing for re-election as a member of the SIG Executive Committee. This announcement was greeted with massive applause. After a break, the meeting broke up into three working groups. The first group, led by Bruce Mitchell, discussed the topic of PDP-11 and RSX SIG futures. The second group, led by Gary Maxwell, dealt with volunteer recruitment and retention. The third group, led by Rick Sharpe, evaluated the present SIG products and personnel. The meeting re-formed after the working groups completed discussion. A summary of their reports follows. PDP-11 and SIG Futures group: It is likely that we have seen the last new architectural implementation of the PDP-11, though new products based on the J11 may yet be built. There will thus be incremental changes, but no major new CPUs. PDP-11 multiprocessing is of course still a possibility. As for commonality with RSTS and RT SIGs, the group felt that the basic missions and needs of the SIGs are not the same. Amalgamation of these SIGs would only dilute their missions. Additionally, any chair of a "16-bit SIG" would naturally feel a loyalty to his own operating system, jeopardizing the future of other groups within the SIG. The group divided SIG goals and mission into short-term and long-term. Short term SIG goals are to support users and RSX coprocessing systems. Long term SIG goals are keeping the SIG viable, supporting conversions to next-generation hardware, and addressing the "real" mission of RSX - process control and industrial automation. Volunteer Recruitment and Retention group: We should request LDEC to support us for volunteer recruitment, since that is their function. Likewise, we should request the NLC to help us support LUGs, since that is the NLC's function, and LUGs need all the support they can get from us. Gary Maxwell will draft the necessary letters. We have problems keeping volunteers active between symposia. The SIG should publish a bulletin to keep volunteers active. Contents of the bulletin should include activities, contact people, and SIG futures. It was agreed that the bulletin will have a one year trial and be entered into the budget as a new activity. Editorship of the bulletin went up for grabs; it was decided that Gary Maxwell will assume the position. Necessity of the bulletin was questioned when DCS is available. The response was that DCS is not available to all our volunteers, and is unsuitable for the limited readership of the bulletin. SIG Products and Personnel Evaluation group: The SIG has two "products". These products are (1) technical assistance and support, and (2) personnel. Under technical assistance and support, the following items were identified and evaluated: .LIST .LE SIG tapes. OK, continue handling as done now. .LE Campground. Unacceptable, more work required. .LE Clinic. Acceptable. .LE Magic. OK, continue handling as done now. .LE Sessions. OK, continue handling as done now. It was pointed out that our room sizes are too small. We need rooms having capactities of 200-250. .LE Working groups. OK, more work required. Working groups have some serious problems which must be addressed; in specific, goals and longevity. .LE Product feedback to DEC. OK, more work required. .LE Menu. OK, but more work required. The menu needs to be more dynamic, and more give-and-take. .LE Multi-Tasker. OK, continue handling as done now. The Multi-Tasker needs articles, as always. Production is OK. Assignment of topics and lengths may help. .LE Session notes. OK, continue handling as done now. There are some severe problems with the long lead time for submission of session notes. .END LIST Under the general heading of "personnel", the following items were identified and evaluated: .LIST .LE Leadership development: OK, continue handling as done now. .LE Recruitment. Acceptable, but more work required. .LE SIG suite: OK, more work required. There are staffing and time problems with the SIG suite. .LE SIG identity: OK, more work required. We need to promote a unique SIG identity to the "outside" world. .LE Store: Unacceptable, more work required. We lost big bucks in the store and are just beginning to recoup them with the mailing boxes. More unique items are needed. .END LIST .SKIP 2 .CENTER Sunday, February 1 Sunday opened with a general discussion of SIG needs. Symposium campground: The Nashville campground will probably be a temporary area like the one at San Francisco. It needs at least two signs, one at the entrance and one inside. The campground should be publicized in Update.daily. Magic sessions: We need special rewards for Magic / Short Notes contributors. Several items were discussed. Bruce Mitchell will coordinate with Jim Hopp. This led to a discussion of a new store item; Al Bennett will contact Judy Arsenault to see about getting a very limited quantity of into the store on a test basis. There will be a programming contest for Magic. Terms will be announced in the Multi-Tasker ahead of time, so members unable to attend will be able to mail in submissions. Welcome letter: The SIG has been asked to write a welcome letter for new members. This is due by March 15th. Bev Kasper will write this. RSX BBS: A phone budget does exist. Dan Eisner will look into modems. Base hardware has been donated. Control should be implemented by sending a user info packet out by U.S. Mail. Software does exist and will be investigated further. Bruce Mitchell and Jim Bostwick are contact points. After a short break, Ed Cetron reported on SIG classified positions. After lunch, Brian McCarthy addressed the group on RSX and hardware futures. The meeting closed after polling attendees for their opinions of the meeting. The following good points were brought out: .SKIP .NO FILL .NO JUSTIFY o##Position descriptions were filled o##Bob Curley's attendance (good friend of the SIG) o##Got a lot done o##SIG definition and purpose o##The group works well together o##The group included recent members o##Leadership problems were dragged into the light .JUSTIFY .FILL The following bad points were also brought out: .SKIP .NO FILL .NO JUSTIFY o##Bob Curley's attendance (preaching to the converted) o##Occasional digressions from the agenda o##Mongolian food aftereffects o##Bugle music o##Noisy meeting room .JUSTIFY .FILL