.ap .PS 60,72 .LM 9 .RM 70 .NHY .LITERAL #####March Vol. 2 No. 1 1980 .end literal .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 Editor's Comments .s 1 This year the VAX sig tape is being distributed via the LUGS. If you don't have a LUG, start one. You'll enjoy doing business with DECUS. The following LUG's have a copy of the VAX tape from Fall Decus '79. Contact your nearest to arrange acquisition of your very own copy. .s 1 .literal DECUS Mexico LUG Caltech-JPL LUG Jose R. Cen-Zubieta Robert Holzman Le Colegio de Mexico Jet Propulsion Laboratory Camino Al Ajusco 20 MS 125-104 Mexico 20, D. F. 4800 Oake Grove Drive Mexico Pasadena. CA 91103 .eli .tp 6 .literal Bay Area PDP-11 LUG BAYLUG Stan Osborne George Hamma 1678 48th Ave Synergistic Technology San Francisco, CA 94122 20065 Stevens Creek Cupertino, CA 95014 .eli .tp 6 .literal San Diego Commercial LUG Naval Weapons Center RT-11 LUG Richard Baldwin Michael Levine N. County Computer Naval Weapons Center Services Code 3513 2235 Meyers Ave. China Lake, CA 93555 Escondido, CA 92025 .eli .tp 6 .literal LLLLUG San Diego PDP-11 LUG Jim Duffy Edward Mueller Lawrence Livermore Labs Simpact Associates PO BOX 808, L-464 5520 Ruffin Road Livermore, CA 94550 San Diego, CA 92128 .eli .tp 6 .literal Colorado RSX-11 LUG Chesapeake Area RT-11 LUG Art Kuczek Henry Dardy AMAX U. S. Naval Research Lab 5950 McIntyre Road Code 8133 Golden, CO 80401 Washington, DC 20375 .eli .tp 6 .literal Southwest Idaho LUG Indianapolis RSX/VAX LUG Roger Engleman Marem Sanghvi USDA-SEA-AR Indianapolis Center for 1175 S. Orchard Advanced Research Boise, ID 83705 Fortune Fry Ultrasound Lab Room A32 Indianapolis, IN 46202 .eli .tp 6 .literal Kansas LUG Pioneer Valley LUG Leon McCormick Bob McCormick RUF Corp. Farm Credit Banks of Springfield 13700 Pflumn Road PO BOX 141 Olathe, KN 66061 Springfield, MA 01101 .eli .tp 6 .literal MIT VAX LUG RSX-11 Washington DC LUG Wendy Koenig Mary Ann Esfandiari Bates Linear Accelerator NASA/Goddard Space Flight Center 77 Mass. Ave. Code 664 MIT High Energy Physics Lab Cambridge, MA 02139 Greenbelt, MD 20771 .eli .tp 6 .literal Southeastern Michigan LUG St. Louis LUG James Cutler Richard Aurbach U-M Space Physics Lab Monsanto Agricultural Products 2455 Hayward Company Ann Arbor, MI 48105 800 N. Lindbergh St. Louis, MO 63166 .eli .tp 6 .literal Montana LUG Midlands Local DECUS Users Group John Barr Thomas Bridge University of Montana Norco Mills of Norfolk Computer Center P O BOX 56 Missoula, MT Norfolk, NE 68701 .eli .tp 6 .literal Upper Valley LUG New York Metro LUG Steve Ladd John Runyon Creare Inc. 39 Locust Point Road Box 71 Locust Point, NJ 07760 Hanover, NH 03755 .eli .tp 6 .literal PDP-11 Seaboard LUG LASL Users of RSX Mark Diebert Dennis Perry E. R. Squibb Los Alamos Scientific Lab 25 Kennedy Blvd. MP-1, MS-828 East Brunswick, NJ 08816 Los Alamos, NM 87545 .eli .tp 6 .literal Ithica Minicomputer Users Central Ohio VAX LUG Group, Richard Cochran R. Whyde Cornell University Battelle Columbus Labs 301 Thurston Hall 505 King Avenue Ithaca, NY 14853 Columbus, Oh 43201 .eli .tp 6 .literal WRUG Western Reserve Users COPEUS Group: Douglas Fair George Atkins Ohio Turnpike Commission Southwestern OK State University 682 Prospect Street Computer Center Berea, OH 44017 Weatherford, OK 73096 .eli .tp 6 .literal Portland Oregon PDP-11 LUG Carolinas LUG Bill Bagley John Scogin Stan Wiley Inc. Realtors Lockwood Green Engineers 200 S. W. Market Suite 1800 PO BOX 491 Portland, OR 97201 Spartanburg, SC 29304 .eli .tp 6 .literal East Tennessee PDP-11 LUG Tennessee LUG Kenneth Cross Jerry Kiestler Oak Ridge National Labs University of Tennessee PO BOX X Bldg 3500 at Martin Oak Ridge, TN 37830 Computer Center Martin, TN 38238 .eli .tp 6 .literal West Texas LUG Alamo PDP-8/11 LUG William McKee Douglas Threatt United Coupon Clearing House Data Sciences Division 1516 53rd St. SAM/BRS Lubbock, TX 79412 Brooks AFB, TX 78235 .eli .tp 6 .literal North Texas PDP-11 LUG Seattle Area LUG John Jenkinson Raymond French Mostek Corporation Boeing Commercial Airplane Co. 1215 West Crosby Rd. PO BOX 3707 Carrollton, TX 75006 MS-28-60 Seattle, WA 98124 .end literal .s 2 This issue is just a little late to allow the inclusion of an SIR ballot and a couple of questionnaires. These are toward the end of the issue and I hope you each take the time and the postage required to respond. Beings as how this issue is already a bunch late, I will dispense with the witty and brilliant editorial column and let the issue speak for itself. Enjoy! .s 2 .tp 6 Editor, Pageswapper: I learned of your existence only last week when a fellow employee (who is not directly concerned with the VAX) loaned me Vol.1 no.1, 2, and 3 of the PAGE SWAPPER. Perhaps this lack of publicity is the major cause of lack of contributions. We have had a VAX since January of this year. The problem of a rewinding magtape hanging the system has plagued us all this time. Inquiries to DIGITAL did not reveal that the critical circumstances (pressing on-line during rewind) could easily be avoided, so it appears that DIGITAL software personnel are not getting or reading the PAGE SWAPPER. A problem mentioned in Vol. 1 no. 2 of having only one set of attributes for each UAF entry is also an annoying problem at our site. I also suffered hours of grief with SYSGEN until I realized that I needed an explicit "use current", and I agree wholeheartedly with the user who pointed out the uselessness of the default parameter column. We had problems with incompatibility between microcode and versions of floppies. I think there is still a disagreement between field service and RDC about which floppies should be left on-site. I would appreciate hearing more about a magtape which can make a complete copy of an RP06 in 12 minutes (Vol 1 No 3). we expect to purchase a TEU77, which is rated at 125 ips, but I would expect an RP06 to require a minimum of 7 tape volumes at a bare minimum of 4 minutes per volume. We received VMS with a cluster factor of 11, but no warning that this might not be appropriate. Our system RM03 pack quickly filled up. With a little butchery of [SYSUPD]VMSKITCPY and some additional sweat to handle user files, I was able to copy the pack to another with a cluster factor of 6, making over 13% more free space. If I had a clearer understanding of the trade-offs involved and the interaction between cluster factor and extension, I might be tempted to try to free up more space. I could be very enthusiastic about a TECO with support for non-VT52 type terminals, if it includes terminals such as the Soroc IQ120, which does not have a command to clear from cursor to end of screen. many of our users now do screen editing on a single-user HP2100 system (when it's available) and are not nearly as happy to use SOS. A better backup would be even better than a better mousetrap. I agree with the consensus that we would be best served by a system which is "intelligent, on-line, and file structured," albeit slow. We have not yet managed to get DECNET working to allow file transfers between the VAX and RSX-11M in an 11/35. The problem seems to be in the 11M software. We have not kept a chart of our field service experience, but during the first 3 months we operated on the expectation that the VAX would crash at least once daily. After a patch was finally provided for the RM03 driver, the situation improved dramatically. But I cannot yet say I am satisfied with the reliability of the system. The RDC is a real plus and I am pleased with their responsiveness. Yet I must report that on two occasions their diagnosis seemed unrelated to the repair which brought the system back to life. In the early months of '79 I received from DIGITAL 3 short papers on VAX performance, configuration and tuning, including the one by Hank Levy in Vol 1 No 3 of the PAGE SWAPPER. All came from the Maynard/Merrimack support group and a publication called "THE SMALL BUFFER," and were dated from 1978. Sadly, this source seems to have dried up. [Editor's note: dried up just doesn't describe this total drought.] Perhaps the most deficient area of documentation is tuning for optimum response with a given configuration. I have discovered some potentially useful information on PAGEFILE sizing, which i am including in a separate file. Unless you consider portions of this letter appropriate, I have no other contributions ready at this time, but i will continue to write. .tp 10 .s 1 Vanya Matzek .br Measurex Corp. .br One Results Way .br Cupertino, California 95014 .s 1 .c 72;PAGEFILE SIZING The VAX Software Installation Guide (section 4.3.1) says that the space which must be preallocated in [SYSEXE]PAGEFILE.SYS is "the size of the average program in pages multiplied by the number of processes." This presumes a lot: you must be able to determine this average size for a dynamic mix of applications, and you must be able to guarantee that peak loads do not exceed the average. LOGOUT/FULL does show the peak virtual size, but only to each user, not the System Manager. Other factors must also be considered: PAGEFILE usage for a given process should only be virtual size less working set size; peak virtual size may vary considerably during a long interactive session. The info saved in the accounting file appears to be only PAGEFILE use (not total virtual size), but it also consists of only long-term peaks, and there is presently no accounting report program to extract the relevant data and present it in a useful form. If you have plenty of spare disc space it is probably wisest to substantially overallocate, but if you must pay for this space you would probable like some method of determining how much of this file is actually used. The VAX/VMS System Programmer Workbook (pag-30) [Editor's note: What is this Workbook? How do we all get one?] gives a little more info in the form of a figure describing the PAGEFILE data-base, but does not quite tell enough about how it is used. The last two longwords of the data-base contain count-1 of pages which may be allocated and pages which may be reserved. Exec microfiche #5 contains a PAGEFILE module, but it only contains code referencing the data-base thru global symbol MMG$A__PAGFIL. knowing the absolute address of module pagefile from SYS.MAP, it is possible to find the address of the data-base table using the hex code shown in the fiche. While this is an interesting exercise in native-mode instruction decoding and hex arithmetic, the map shows the data-base to reside in module SWAPFILE (fiche #6) at %X800031D8, with the count-1 of allocatable pages at %X800031F0. While the system is running, the values in this table can be monitored with the DCL EXAMINE command (EX/DEC is most useful). Procedure PAGFILCHK.COM can be called with any non-null P1 to determine current free space in PAGEFILE, to maintain the minimum such value in system logical name SSW__MFPF, and to display these values. A call with no parameters displays only if a new minimum occurs. A simple procedure such as SENTRY.COM may be submitted to execute PAGFILCHK periodically. The desired info is readily available in SSW__MFPF and the log file gives a record of fluctuation with user load. M$ is a logical name defined in our SYSTARTUP.COM as device and directory containing managerial procedures such as PAGFILCHK.COM. The installation guide recommends a PAGEFILE of 16384 blocks for a 16 user system, but I found we were using less than 2000 blocks. I have since reduced the PAGEFILE to 9000 blocks (for 16 users) and still see a useage of less than 2000. This is no doubt due to having 1.5 megabytes of CPU memory installed, giving us a large free page list, but now we can monitor the PAGEFILE space as additional user terminals are added. .tp 8 .s 1 .literal $! PAGFILCHK ONE-SHOT CHECK PAGEFILE.SYS FOR FREE SPACE $! INITIALLY $ASSIGN/SYSTEM 14000 SSW__MFPF !MINIMUM FREE IN PF $! 800031F0 (VMS 1.5) IS EXAMINED & USED TO UPDATE MINIMUM $! P1.NES."" FOR DEBUG, SHOW CURRENT $SET NOON $SET NOVERIFY $MFPFS:='F$LOG("SSW__MFPF") $MFPF="''MFPFS'" !CURRENT MINIMUM $IF MFPF .EQ. 0 THEN MFPF=14000 !DEFAULT, RESET $OPEN/WRITE COMMAND EXPAGE.TMP $WRITE COMMAND "$EX 800031F0" $CLOSE COMMAND $SET NOVERIFY $@EXPAGE.TMP/OUT=X.TMP $DEL EXPAGE.TMP;* $OPEN/READ VALUE X.TMP $READ VALUE LINE $CLOSE VALUE $DELETE X.TMP;* $CFPFS:='F$EXTRACT(11,8,LINE) !GET VALUE STRING $CFPFHS:="%X''CFPFS'" !MAKE HEX STRING $CFPF='CFPFHS !CONVERT TO DECIMAL $IF CFPF .GE. MFPF THEN GOTO AT__MIN $! ELSE SET NEW MINIMUM VALUE $SHOW TIME $ASSIGN/SYSTEM 'CFPF' SSW__MFPF $SHO LOGICAL SSW__MFPF $SET LOGIN/INTER !SHOW NUMBER OF USERS $ AT__MIN: $IF P1 .NES. "" THEN SHOW SYMBOL CFPF $IF P1 .NES. "" THEN SHOW LOGICAL SSW__MFPF $EXIT $! SENTRY.COM MONITOR SYSTEM PERFORMANCE $! $SET NOVERIFY $ LOOP: $WAIT 00:01 $@M$:PAGFILCHK $GOTO LOOP .end literal .tp 8 .S 2 ANNOUNCEMENTS .s 1 In keeping with our continuing policy of disseminating information about interesting software for VAX, the following section will contain product announcements and information. There will be no price and delivery information, just 'what's available'. .s 1 Tape Driver:-Interfaces Storage Technology Corporation (STC) tape drives to the VAX. Handles 125 ips, 9-track (1600, 6250 bpi) and 7-track (200, 556, 800 bpi) drives. .s 1 XEROX 1200:- A queue manager is now available to handle dequeueing for a XEROX 1200 Off-line Printing System. .s 1 SORT:-An efficient sort/merge program written in FORTRAN, handling fixed and variable length records. Requires maximum work space of 2.5 times the size of the input file (current DEC sort requires 6 to 8 times the size of the input file) and will accept a user collating sequence with up to nine keys of the following types: .list .le VAX integer (1,2 and 4 bytes) .le VAX floating (single and double precision) .le ASCII .le EBCDIC, BCD and PACKED (IBM) .le Control Data Integer and Floating Point .end list .s 1 TRANSACTIONAL RECORDING:-A system owned mailbox manager and a user interface utility provide quick and secure delivery of free-format transaction records independent of regular system accounting functions. .s 1 The above four packages are available from MSA in Pittsburgh. Contact Mark Pilant at MSA, Shadyside Centre, 5100 Centre Avenue, Pittsburgh, PA 15232. .s 2 UNIX WORKBENCH:- The Workbench system on VAX/VMS was developed to provide UNIX (trademark of Bell Laboratories) system functionality under the VMS operating system. It is comprised of the following major subsystems: .lm 14 #1. C compiler .br #2. Unix system primitives .br #3. Standard I/O primitives .br #4. Programmer's workbench primitives .br #5. Shell command interpreter .br #6. yacc - compiler-compiler .br #7. lex - program generator .br #8. m4 - macro processor .br #9. nroff - text formatter for terminals .br 10. tnroff - text formatter for phototypesetter .br 11. eqn/neqn - mathematical equation formatter .br 12. tbl - table formatter .br 13. SCCS - source code control system .br 14. Make - program maintenance facility .br 15. INed - screen editor .br 16. INmail - message system .br 17. INword - word processing system .br 18. many other unix programs .lm 9 .s 1 One of the major goals of this software development effort was to provide all of the UNIX software in the VAX/VMS environment in a compatible fashion. These Programmer's Workbench tools were provided to facilitate the development of native mode VMS programs. Files are compatible between the UNIX and the VMS programming environments. Programs may be called either from DCL or from the shell command interpreter. In fact many 'shell' features are available directly from DCL. The C compiler calling sequence is compatible with the VMS routine calling sequence. Most of the UNIX programs now running under VMS were originally written for the UNIX system on the PDP-11 computer. A result of this effort has been a large set of portable programs which run both on the VAX and the PDP-11 computers. For information contact Heinz Lycklama, Interactive Systems Corporation, Santa Monica 213-450-8363. .tp 10 .s 2 To: VAX SIG .br From: Ben Mossberg .br Dept. of Civil Engineering, University of Southern California .br University Park, Los Angeles, CA 90007 .br 213 741 5809 .s 1 Subject: ACT 2.2 VAX/VMS Accounting System .s 1 After one year of experience with this package on our VAX we have had excellent success. This package controls user access and maintains user accounts. It gives the system manager good control over usage, access and funds allocation, along with some very useful summary reports. ACT 2.2 is one of those rare notable programs where no bugs or problems have been discovered dring last year. If you need some usage accounting on your VAX, I highly recommend this package. It is available under perpetual license. If you are interested, give me a call. .tp 6 .s 2 VAX/VMS SPR's of Interest .s 1 [Editor's note- In this column we wish to present SPR's that are useful, informative, interesting, aggravating... The purpose is to share common problems and solutions. This column is also designed to fill in the gaps left by the absence of a VAX/VMS Software Dispatch. Another purpose is to disseminate DEC's SPR responses which solve problems for many users. The final purpose is to get real solutions to SPR's which DEC has closed without solving. Send me copies of SPR's (and responses if approporiate) and I'll try to put the essence into this column. Please be aware that the folks who respond to SPR's are willing to review their responses when we, the users, feel that they are inadequate. The mechanism to activate this review is a second SPR. This SPR should reference the number of the original SPR, redefine the problem, indicate that the solution presented in the initial response is inadequate, and give details as to why this solution is inadequate.] .s 1 .tp 4 ------------------------------------------------ .br When multiple PARAMETER statements define the same constant, the error message %FORT-F-ERROR#20, Missing variable or subprogram name is generated. This is very misleading. .s 1 Response .s 1 Thank you for noting this error. It will be corrected in the Version 2 release of VAX-11 FORTRAN. .s 1 .tp 4 ------------------------------------------------ .br DEBUG does not display the CASE statement displacement word if the limit operand has value zero (i.e. when there is one displacement word). This causes the current PC to be off by two bytes. This in turn causes garbage to be printed with an EXAMINE/INSTRUCTION is used to look at the next instruction. .s 1 Response .s 1 Thank you for your SPR. The problem has been corrected in Version 2 of DEBUG-VAX. .s 1 .tp 4 ------------------------------------------------ .br It would be nice to be able to put strings (either character strings or hollerith constants) in DATA statements without breaking them into 4 character substrings: .literal DIMENSION ISTR(2) DATA ISTR/8HHI THERE/ .END LITERAL Response .s 1 This capability is already present in VAX-11 FORTRAN. For example: .literal CHARACTER*8 MSG DATA MSG/'HI THERE'/ .END LITERAL See sections 5.2.2 and 5.7 of the FORTRAN language reference manual. .s 1 THE ORIGINAL SPR WRITER RE-SPR'S AS FOLLOWS: .s 1 You completely missed the point of the SPR. The CHARACTER variable type is an altogether different data type. I was specifically referring to non-character data types: integer, real, byte, etc. This would help in the conversion of programs from other machines to the VAX. .s 1 Response .s 1 Thank you for the suggestion. It was not stated in the previous SPR that only numeric data types were being considered. If this issue continues to be raised by the user community we will consider including it in a future release of VAX-11 FORTRAN. This suggestion has been considered before and rejected on the grounds that conversion of string data to CHARACTER type is a more productive activity given the existence of the new FORTRAN-77 standard. .s 1 In the meantime, note that with suitable equivalence to byte arrays, you can accomolish what you desire today, such as: .s 1 .literal INTEGER*4 ISTRING(6) BYTE BSTRING(24) CHARACTER CSTRING*24 EQUIVALENCE (ISTRING,BSTRING),(BSTRING,CSTRING) DATA CSTRING /'NOW IS THE TIME FOR ALL'/ .END LITERAL .S 1 .tp 4 ------------------------------------------------ .br JOB CONTROLLER does not always deallocate the device when the queue is stopped. .s 1 Response .s 1 Thank you for reporting this. If the STOP/QUEUE/NEXT command is issued when the queue is idle (no printing in progress), the device will be deallocated. If this command is issued while there is printing in progress the device will not be deallocated. This is because the JOB CONTROLLER attempts to deallocate the device immediately upon receiving the command. The deallocate fails because the print symbiont still has a channel assigned to the device. This problem will be corrected in Release 2 of VMS. .tp 4 ------------------------------------------------ .br The PURGE utility does not work of there are more than approximately 1000 to 1200 files in a directory. No purging is done and an error message is returned: "DELINT ERR", INTERNAL ERROR IN DELETE/PURGE UTILITY. .s 1 Response .s 1 This is caused by the fact that purge reads the entire directory file into the program region in order to determine which files should be deleted. If there is not enough space, an error results causing the message you see. This will be fixed in release 2.0 of VMS. Until then, use PIP/PU on those directories that fail. .tp 4 ------------------------------------------------ .br When the .NTYPE operand argument is the PC register, an 'ILLEGAL REGISTER USED HERE' error message is always generated. .s 1 Response .s 1 This problem will be fixed in the version of MACRO-32 included in VMS 1.7. .tp 4 ------------------------------------------------ .br In a listing of a MACRO-32 program the address displayed for a MOVA or a PUSHA wi the relocatable address of the symbol and not the offset that is put in the object file. .s 1 Response When the MOVA or PUSHA instruction and address are in the same PSECT and the assembler can calculate the offset, an offset value is displayed in the assembler listing. Otherwise, the assembler must output a relocatable address to the object module so that the linker can calculate the offset when the positions of all PSECTS are known. In this case the assembly listing shows this relocatable value. .s 1 .tp 4 ------------------------------------------------ .br It would be useful is user defined register names could be created, and that MACRO-32 have predefined rather than reserved symbols. .s 1 Response .s 1 Your suggestion, for user defined register names, will be considered for inclusion in a future release of MACRO-32. .s 1 Reserved symbols (opcodes, directives, etc.) can be redefined by using macros or the .OPDEF directive, in which case the new definitions override the built-in functions. .s 1 .tp 4 ------------------------------------------------ .br When pressing return whil the curson is in column 0 a line feed isn't always echoed. .s 1 Response .s 1 The terminal driver should echo CR/LF when CR is typed, this problem will be fixed in V1.5 of VMS. .s 1 .tp 4 ------------------------------------------------ .br The index of an implied DO loop in a READ/WRITE statement is not properly materialized if the statement terminates using an END= or ERR= option. A short term solution is to initialize the implied index to the starting value before executing the READ/WRITE statement. The apparent long term solution is to fix the compiler. .s 1 .tp 4 ------------------------------------------------ .br __%LOCATE does not work in a nested MACRO. .s 1 .tp 4 ------------------------------------------------ .br The LIB$INITIALIZE initialization co-routine (Appendix E of Common Run-Time Procedure Library Reference Manual) is actively defeated by FORTRAN STOP statement. .s 1 .tp 4 ------------------------------------------------ .br In SYSGEN parameter TTY____DEFCHAR bit TT$V____REMOTE seems to have no effect. One result of this is that for a terminal to be eligible for REPLY/ENABLE an explicit SET/TERMINAL/LOCAL must be done. .s 1 Explicit documentation would have made tracking this problem much easier, but we consider the impotence of this bit to be a problem in any event. .s 1 .tp 4 ------------------------------------------------ .br The inability of VAX DECNET to accept a Null password per the standards called out in NSP Functional Specification Version 3.1 section 4.2.4.2.2. This manual requires communicating RSX11D systems to need manual intervention. This results in an automatic re-boot (i.e. following power failure). After re-boot the manual workaround is to enter passwords on all RSX11D systems in the network. This is done to overcome the inability of VAX to accept the "DECNET standard" of NULL passwords. The DEC suggested fix of setting up a BATCH QUEUE on the RSX11D systems is totally our of the question in our application and does not address the heart of the problem. .s 1 Response .s 1 Although NSP permits accepting nell node initialization passwords, it does not require accepting them. As you stated, DECNET-VAX requires non-null passwords. However, DECNET-VAX does conform to the NSP standard. .s 1 The inability of DECNET-11D to send non-null node initialization password conveniently under reboot conditions should be the subject of an SPR to DECNET-11D. This SPR will be forwarded to the DECNET-11D maintenance group. In addition we are considering accepting null node initialization passwords in the next release. .s 1 .tp 4 ------------------------------------------------ .br After a FORTRAN direct read (with ERR= specification and fixed length records) that fails, WRITES may also fail with RMS error RSZ indicating that the record size field was invalid. .s 1 Response .s 1 Thank you for finding this problem. The Run Time Library was unaware that RMS could change the RSZ field for fixed length records on $GET's. In fact, RMS sets the record size field to zero for certain error conditions. This causes future calls to RMS which depend on RSZ to also fail. This will be fixed in release 1.6 of VMS. .s 1 .tp 4 ------------------------------------------------ .br The /RC switch causes RST to fail and the error message "RST - check After Write Error on Output File..." to appear under the following conditions. .list .le BCK is used to create a tape from files on an RP06 where the allocation cluster size is 11. .le The tape is used to restore onto an RP06 with an allocation cluster size of 3. .end list It appears that RST requests an initial allocation equal to the allocated size of the old file (not the used size), writes the file, and then compares the entire allocated space of both files, thus generating the above described error. .s 1 It would seem appropriate in any fix to: 1) make RST request allocation equal to the used figure (i.#e. pay attention to EOF) and 2) when in /RC mode, check only to EOF, not to the end of allocation. .s 1 .tp 4 ------------------------------------------------ .br We have recently reconfigured all our disks to an allocation cluster size of 3 and a default extension size of 6. Previously these values were 11 and 5 respectively. We find now that the volume default extension value does not work. As the enclosed listings show, the disk home block has been properly initialized with the desired values, but a FORTRAN program that should cause the default extension to occur gets only an additional cluster of 3 blocks. In addition, the SOS editor allocates 6 blocks initially and extends by 6. However if the files are subsequently edited, the allocation drops to "used" rounded up to the next larger integral cluster. [Editor's question: Is this cluster the default volume cluster?] .s 1 .tp 4 ------------------------------------------------ .br .tp 10 .s 2 Unsolicited Random Input .s 1 [Editor's note: In this section we will present interesting trivia and helpful facts about VAX/VMS. The sources will be unnamed at present to protect the innocent.] .s 2 The following VAX/VMS wish list has been presented to DEC; the items on the list are ranked in priority order. .list .le A system maintained "last access field" in the disk file header. Updated on any open, this field is required for any rational file cleaning algorithm. An additional "times used since creation" field would also be highly useful. .le A reasonable scheme for enforcing maximum disk space allocated per user. We have heard of a "guillotine" scheme in Version 2 which is unacceptable. In theory, a user should have dynamic and unlimited disk space during a process lifetime. Only at logout or login should disk quotas be enforced. [Editor's note: I am not sure I quite agree with the unlimited quotas. In the situations where quotas are necessary at all, unlimited in-process quotas would be immediately abused.] .le A backspace facility within the file system. Rewinding and spacing forward is not the answer to the problem. .le A modified DSC to allow specification of output disk volume space parameters such as cluster size, extend size, label, etc. [Editor's note: Hear! Hear!!] .le A facility for dumping and restoring batch queues. This facility is required when the system disk is changed, as when DSC is performed, and the queues are non-empty. .le "Backspace n pages" command for print symbionts. It is inadequate to have just "top of file" since large print jobs are often involved, and "top of file" wastes paper and time. .le A complete definition of the CLI callback facility and its uses would be very useful for systems work. .end list .s 1 .tp 4 ------------------------------------------------ .br The following are the graphics from Dick Hustvedt's talk at Decus on performance and tuning. .literal Slide 1 HOW TO TUNE VMS FOR OPTIMUM PERFORMANCE o BUY MORE MEMORY Slide 2. MEMORY ALLOCATION o NPAGEDYN-NON-PAGED DYNAMIC MEMORY o I/O DATA BASE o I/O DRIVERS o MEMORY MANAGEMENT o OTHER SYSTEM STRUCTURES o IRP/COUNT - I/O PACKET LOOK-ASIDE LIST Slide 3. PROCESS CONFIGURATION o VIRTUALPAGECNT-MAXIMUM PROCESS VIRTUAL ADDRESS SPACE o MAXPROCESSCNT-MAXIMUM PROCESSES IN SYSTEM o WSMAX-MAXIMUM PROCESS WORKING SET o BALSETCNT-NUMBER OF BALANCE SET SLOTS Slide 4. ACP TUNING o LEAVE ACP____SHARE AND ACP____MULTIPLE OFF o ACP____SYSACC - DIRECTORY PRE-ACCESS FOR SYSTEM DISKS o ACP____SWAPFLGS - SWAP ENABLE FLAGS Slide 5. ACP CACHES o ACP____MAPCACHE - STORAGE BITMAP o ACP____HDRCACHE - FILE HEADERS o ACP____DIRCACHE - DIRECTORIES AND OTHER DATA Slide 6. TRADING OFF PAGING AND SWAPPING o DON'T DO BOTH o LET SMALL JOBS SWAP o SEQUENCE LARGE JOBS o KEEP FILE I/O SEPARATE o TUNE YOUR APPLICATION Slide 7 PAGING SYSTEM o FREELIM - FREE LIST LOW LIMIT o MPW____LOLIMIT - MODIFIED LIST LOW LIMIT o MPW____HILIM - MODIFIED LIST HIGH LIMIT o MPW____WRTCLUSTER - MODIFIED PAGE WRITE CLUSTER o SYSMWCNT - SYSTEM WORKING SET Slide 8 RELEASE 2.0 PERFORMANCE ENHANCEMENTS o ACP I/O REDUCTION o AUTOMATIC WORKING SET ADJUSTMENT o SWAP RATE LIMITING (SWPRATE) o CHANGED QUANTUM CHANGE FOR WAIT TIME (IOTA) o CORRECTLY IDENTIFY TERMINAL I/O o CORRECTLY IDENTIFY ACP I/O Slide 9 FILE ACP I/O REDUCTION o EXTENT CACHEING o FILE ID CACHEING o WRITEBACK FILE HEADER CACHE o QUOTA BLOCK CACHE o NET IMPROVEMENT AVERAGES 50% I/O REDUCTION o NO SACRIFICE OF STRUCTURE INTEGRITY Slide 10 AUTOMATIC WORKING SET ADJUSTMENT o PAGE FAULT RATE THRESHOLDS HIGH (PFRATH) LOW (PFRATL) o WORKING SET MODIFIERS INCREMENT (WSINC) DECREMENT (WSDEC) o SIZE BOUNDS MAXIMUM (AWSMAX) MINIMUM (AWSMIN) o MEASUREMENT INTERVAL (AWSTIME) o OPTIONAL - DISABLED FOR REALTIME .end literal .page .p0,2 THE SIG SYSTEM IMPROVEMENT REQUEST PROGRAM .p0,2 The VAX SIG is sponsoring a program of System Improvement Requests designed to improve communications between the users of VAX systems and the DEC developers. The SIR program is designed to augment the SPR system, and to solicit direct response to user request from the developers. Unlike the SPR, which is intended as a means of reporting bugs or needed enhancements, the SIR is intended as a forum, suggestion box or grafitti board for all areas of interest to VAX (not just VAX/VMS) systems. Suggestions for new hardware support, new utilities, new operating systems??? enhancements to VMS are all appropriate topics for the SIR. .p0,1 Requests for improvements to the VAX system are submitted to the coordinator on a SIR reporting form. These requests are collected, screened for duplicates and compiled into form for printing in the news letter. Between each DECUS Symposium an issue of the Page Swapper will be printed containing all of the SIR's collected for that time period along with a ballot. The user community will fill out their ballots and return them to the coordinator. The most popular of the SIR's reported will be answered directly by the developers in a presentation held at the following DECUS Symposium. The operation of the SIR program is being coordinated by Doug Wilson of The Joint Computer Facility at M.I.T. For obvious reasons the number of SIR's that are formally responded to must be finite; however, the developers have requested that all of the SIR's be forwarded to them for their review. .p0,1 This is the first issue of the Page Swapper to contain SIR's and the ballot form. Sorry that we are late and somewhat disorganized, but we are here! Please review the SIR's enclosed and mark your ballot accordingly. Next time around we hope to have many more for you to look at. We need your input!! As we are going to print I have only received about 20 SIR's. I cannot believe that these are the only improvements that the user group would like to see in VMS!! In the next issue of the Page Swapper we will print a SIR reporting form; with any luck we will also have dial in service (in the northeast any way...we have a secure system we think...any volunteers???). For now just send them in on a piece of paper. Send them to .br .lit Douglass J. Wilson M.I.T. Joint Computer Facility Room 1-106 77 Massachusetts Ave. .ELI .page .PAGE .P-10,2 SIR #: S-80-0001 .P-10,2 ABSTRACT: .br MICROCODING DEBUGGING TOOLS .p-10,2 DESCRIPTION: .br As part of the tools for microcode development, which now includes a microcode assembler, definition file, and loader, there is also a need for some debugging tools. Debugging tools similar to what is available for microcoding the LSI11 would make a good model. A snapshot buffer of the last N microinstructions executed would be useful. .p0,2 I think it will be very difficult to write microcode without debugging tools. .TP 7 .p-10,2 SIR #: S-80-0002 .p-10,2 ABSTRACT: .BR We request a multiprocessing ability for each individual user as is done on RSX. Batch submission is desired, but the above requested capability is also very desirable. .p-10,2 DESCRIPTION: .br Our users are often previous users of RSX and LTSS (a LLL, LASL operating system on 7600's and CRAY's). They are quite astute at using a multiprocessing capability as well as batch. We need this capability to enable us to compete in the eyes of these users. We also need firm ware ie. fast methods of byte swapping relative to 16 & 32 bit fields and blocks of bytes. .TP 7 .p-10,2 SIR #: S-80-0003 .P-10,2 ABSTRACT: .br File header should include 'date-last-accessed' and this field should be updated every time file is accessed. .p-10,2 DESCRIPTION: .br When files have not been accessed for a specific period of time (i.e.: 3 months or 6 months) then they could automatically be migrated to a tape backup and removed from on-line storage. .br This capability is essential to maintaining an efficient operation and use of on-line disk storage. Any decent backup utility should require this field. This is probably difficult to implement but would be well worth the effort. Version 2.0 does not include this date. .tp 7 .p-10,2 SIR #: S-80-0004 .P-10,2 ABSTRACT: .br SORT VERB added to VAX COBOL .p-10,2 DESCRIPTION: .br We have at least 190 COBOL programs using SORT VERB currently on an IBM360/30. Implementation of SORT VERB will allow us to eliminate need for the IBM360/30. .tp 7 .p-10,2 SIR #: S-80-0005 .P-10,2 ABSTRACT: .br RK05 should be supported on VMS. .p-10,2 DESCRIPTION: .br The RK05 is a disk widely used for communication of software. VMS should suport this device as it would facilitate moving information between VAX and PDP-11 machines. .tp 7 .p-10,2 SIR #: S-80-0006 .P-10,2 ABSTRACT: .BR Support the display of parameters associated with a queued batch job. .p-10,2 DESCRIPTION: .br Batch jobs can be submitted with 0-8 parameters. These parameters are often used to tailor generic command procedures. There is at present no means for determining the parameters with which a queued batch job was submitted. This complicates system usage and operation because there is no way of determining what these generic command procedures have been asked to do. For example, a multi-person programming project may do all of its compilations using a single, parameterized command procedure. There is no way to determine which modules have been submitted for recompilation by other programmers. If batch job parameters could be displayed, then this information would be available. .tp 7 .p-10,2 SIR #: S-80-0007 .P-10,2 ABSTRACT: .br Support the cross-reference feature of MACRO-11 and Task Builder. .p-10,2 DESCRIPTION: .br VMS is a very suitable host for RSX-11M program development. One of its few shortcomings is that the cross-reference feature of MACRO-11 and Task Builder do not function under VMS. Cross-reference is a highly desirable feature when large programs are being developed. .tp 7 .p-10,2 SIR #: S-80-0009 .P-10,2 ABSTRACT: .br Support allocating drives and mounting volumes on remote nodes. .p-10,2 DESCRIPTION: .br It is difficult to use private volumes on a remote node because DECNET provides no mechanism for allocating a drive or for mounting a volume. This makes it difficult to use DECNET to share disk and tape drives between systems. .tp 7 .p-10,2 SIR #: S-80-0010 .P-10,2 ABSTRACT: .BR Support error logging on DMC-11. .p-10,2 DESCRIPTION: .BR Using the SHOW DEV command, we observe on occation non-zero error counts for our DMC-11's. There apears to be no way of determining what kind of errors are happening. We suggest that these errors should be logged so that information about them will be available. .tp 7 .p-10,2 SIR #:S-80-0011 .P-10,2 ABSTRACT: .BR Provide an operator with a command to remove a process from contention for memory and a command to reinstate a process previously removed. .p-10,2 DESCRIPTION: .br VMS systems with limited memory may sometimes get into a state where swapping seems to be the primary activity of the system and response to interactive users becomes intolerably slow. If an operator could remove one or more processes from contention for memory (say by forcing it to be swapped out and to remain swapped out), there would be a mechanism for dealing with this problem short of deleting processes. Of course, it owuld be necessary also to be able to reinstate a process that had been forced out of memory. .tp 7 .p-10,2 SIR #: S-80-0012 .P-10,2 ABSTRACT: .br Auto-logout of inactive terminals. .p-10,2 DESCRIPTION: .br It would be helpful if VAX/VMS included a SYSGEN parameter which, when set, would cause an interactive process (terminal) to be terminated (logged out) if no I/O occurred after a specified time interval. This is useful in installations which charge users for connect time, and where users occasionally remain logged into the system overnight because they forgot to logout. Note: IAS includes this feature. .tp 7 .p-10,2 SIR #: S-80-0013 .P-10,2 ABSTRACT: .BR The concept of file ownership as distinct from file location is generally annoying adn only occasionally useful as imlemented. If ownership were defaulted to directory of location, unless EXPLICITLY named, things would behave more naturally. .p-10,2 DESCRIPTION: .br It is quite common to find files owned by [a] in the directory of [b], and therefore inaccessable to [b]. This usually happens when [a] attempts to 'give' the file to [b] by using $ COPY. Often the intervention of the system manager is required. Many damaged keyboards would be avoided if, upon file creation, the ownership would DEFAULT to that of the directory or subdirectory wherin the file is being created. Strange combinations could still be provided for by /OWNER=______. .tp 7 .p-10,2 SIR #: S-80-0014 .p-10,2 ABSTRACT: .br Subprocess Improvemnets .p-10,2 DESCRIPTION: .br In the area of subprocesses, a number of things have been cleaned up, including the quota assignment problem. However, there are some other areas that need improvement. A subprocess should have more information about its parent process availible. More logical names should be passed from the parent process to it's subprocess. As much of the parent's environment as possible should be passed along to the subprocess. There should also be the notion of a 'home' directory. This should be available to the subprocess as well as the parent process directly by means of a logical name translation. The idea of a process group would be a nice concept to propagate in future releases of VMS. I believe you have used this to advantage in making quota assignments on the basis of a process being a member of a process group. Similarly this could be used to make much more information availible to subprocesses. Each member of a process group should have access to certain common tables shared by all members of that process group. This would make the passing of logical name tables between related processes a natural extension of the current implementation. .tp 7 .p-10,2 SIR #: S-80-0015 .p-10,2 ABSTRACT: .br RMS should support STREAM files .p-10,2 DESCRIPTION: .br In our implementation of UNIX on VAX/VMS, one of the major inefficiencies occurs as a result of having to use the RMS file subsystem to read/write all files. The RMS system imposes unnecessary structure on the files. As you know the UNIX file system treats all files as a simple stream of bytes. This makes it easy to have device independence and to implement efficient file I/O. Is there any possibility at all that a simple UNIX-like 'stream mode' file will be introduced in the VMS system? This notion has been brought up previously in discussions with members of the ARPA community. To make this work in a consistent manner, all current VMS programs would have to recognize this type of file and treat it as a sequential file. Each 'record' should be terminated by a NEWLINE character. .tp 7 .p-10,2 SIR #: S-80-0016 .p-10,2 ABSTRACT: .br File status and protection .p-10,2 DESCRIPTION: .br In dealing with files, there are currently some deficiencies in the allowed operations and in the displayed status of the file. As an example, when the status of a file is dispalyed, the type of file should be indicated clearly. There should be mode bits for each of a sequential file, random access file, a UNIX-like stream file, a directory, etc. One of the glaring missing features now is that a directory is not distinguished from other files in the status bits displayed with the file. The fact that a file is a directory is now only indicated by the name of the file. .p0,2 To change the protection bits of a file, one needs to have write permission on the file if the protection bits are attempted to be modified under program control throught the RMS system. This is an unnecessary protection. If one ownes the file, one should be able to set whatever protection bits one desires. Currently, one can only do this from DCL. One minnor suggestion. It is currently possible to refer to a particular version of a file by means of file.c.3 for example, as well as file.c;3 . Will this continue to remain possible? It is a convenience to thise who are familiar with the shell command language syntax. The shell treats ';' as a special character. .tp 7 .p-10,2 SIR #: S-80-0017 .p-10,2 ABSTRACT: .br DCL Improvements .p-10,2 DESCRIPTION: .br The design of the DCL interpreter leaves something to be desired. For example, DEC advertises that is is possible to the user to write his own command interpreter. IN fact, that is only true if the user's command interpreter fits in less than 48 pages of memory. The number 48 is built into the VMS operating system, where in fact it should be either sysgenable or a variable set by the actual loading of the commend interpreter. In our case, we would like to combine the best features of DCL and the UNIX shell to make a very flexible command interpreter for the user. With the current design of VMS and DCL, this is impossible to do. .p0,2 In making future extensions to the DCL language, DEC should consider the following features. There must be more flexibility in the user command interface. For example, any user command typed in should be treated first as a logical symbol and then as an executable inage to be run automatically, without having to type 'run image' explicitly. One should not have to install the command as a foreign command in order to call it from DCL. In fact there should be a 'search path' associated with the command. If the image is not found in the current directory, it should be searched for in the user's personal home directory and if not found there, in the system directory SYS$SYSTEM. Moreover, the order of the search should be specifiable by each user. This would go a long way to giving DCL some of the power of the UNIX shell command language. The DCL syntax should allow the creation of asynchronous communicating processes, running in the background. All of the DCL translations and interpretation of arguments should be removed for foreign commands. DCL can recognize immediately if the given command is one of the internal ones to DCL. If it is not, lower case characters should not be mapped to upper case character, etc. Many of the modern control structures should be added to DCL, These include 'if-then- elseif', 'while', 'do', 'for', etc. The person who designs new features into DCL should really take a close look at the UNIX shell command syntax and add some of it's features. .tp 7 .p-10,2 SIR #:S-80-0018 .p-10,2 ABSTRACT: .br Terminal driver fixes/enhancements .p-10,2 DESCRIPTION: .br The terminal driver should have several problems fixed. Consistent handling of Write attention AST's is important. On read with timeout...the driver should return to the user all of the characters which have been typed up to the timeout. The current implementation drops all of these characters. .p0,2 The driver should be enhanced with a mechanism to deal with cursor control for strange terminals. Perhaps this could be done with a terminal type field that would cause the driver FDT routines to execute a user written sharable section for standard (vt100/vt52) escape sequences. For users with many dial in lines it would be quite nice if the terminal driver could do auto speed select. Syntax would need to be added to SET/TERM to specify an alternate speed for the terminal. one of the RS232 secondary characteristic line could be used for this purpose. I know of one VAX installation who has done this with a DZ already. .tp 7 .p-10,2 SIR#:S-80-0019 .p-10,2 ABSTRACT: .br Display utility enhancements .p-10,2 DESCRIPTION: .br The Display utility should provide some mechanism for caputuring the output on paper. It is difficult to monitor the behavior of the system over an extended period of time.or with a short display cycle time. An option in DISPLAY to write to a disk file would partially alleviate this problem. A program to produce a printed report from such a file would also be nice. .tp 7 .p-10,2 SIR#:S-80-0020 .P-10,2 ABSTRACT: .br Display utility enhancements .p-10,2 DESCRIPTION: .br The display utility should be able to work on more terminal types. While it is unreasonable to expect DEC to support all vendors terminals, it is perhaps possible that display be built with a shared section that could be replaced by a users written shared section that does the appropriate cursor control functions. .p0,2 The documentation needed would be the functional specs for the cursor control package. .tp 7 .p-10,2 SIR #:S-80-0021 .p-10,2 ABSTRACT: .br NAMELIST I/O Should be included in FORTRAN .p-10,2 DESCRIPTION: .br The inclusion of NAMELIST I/O to the FORTRAN compiler is urgently needed. The computing done here is mostly scientific and involves codes with many variables of which many can be left with default values. These codes which have been running on our two DEC-10's a CDC-7600, CDC-CYBER-172, and IBM-370/158 use the NAMELIST feature to solve the above problem. Without NAMELIST our conversion to the VAX is severely hampered and in our view NAMELIST is a necessity. .tp 7 .p-10,2 SIR #:S-80-0022 .p-10,2 ABSTRACT: .br User extensible CLI .p-10,2 DESCRIPTION: .br The CLI should be able to be user extensible. The system manager should be able to add new commands and parsing tables to DCL in order to add new commands. Also it may be desirable to remove some commands from the CLI. Documentation on the CLI callback facility would also help. We have found that many complex functions can be done efficently from a program if the CLI callbacks are used. .tp 7 .p-10,2 SIR #:S-80-0023 .p-10,2 ABSTRACT: .br Execute only privilege for command files .p-10,2 DESCRIPTION: .br It is desirable the users be able to execute command procedures without being able to read them. This prevents users from "exploring" and finding obscure programs that they sould not execute without benifit of the command procedure. This is particularly desirable in a "hackig prone" environment such as ths academic community. .tp 7 .p-10,2 SIR #:S-80-0024 .p-10,2 ABSTRACT: .br Enhanced provilege for command files .p-10,2 DESCRIPTION: .br There should be some mechanism for "installing" a command file with elevated privelege. This would allow system command procedures to be run by otherwise unprivileged users. Great care must be taken though so that there is no way to escape from the command file with privilege. .tp 7 .p-10,2 SIR :#S-80-0025 .p-10,2 DESCRIPTION: .br DECNET/VAX should support an ASYNCH device .p-10,2 ABSTRACT: .br DECNET VAX desperatly needs an Asych device like the DZ11. For communication with any small systems (LSI 11/23, PDT, ect) Asynch is the only way to go. In addition the cost of synchronous lines modems DMC11's !! is prohibitive for a lage laboratory network with a low duty cycle. With the Sychronous limitation Networks will not be set up for convienence but only for necessity. .page .literal VAX Special Interest Group SYSTEM IMPROVEMENT REQUEST PROGRAM BALLOT Instructions: In the spaces below, please list, in order of importance to you or your installation, the system imporvements that you would like to see implemented on/in VAX Systems. Line 1 should be the number of the SIR you consider most important (or desirable). Line 10 si the least important. Remove the ballot page from your Page Swapper, fold in thirds, staple closed and mail. Please return ballots promptly to allow DEC time to prepare responses. RANK OF IMPORTANCE SIR NUMBER # 1 (most important) _______________________________________ # 2 _______________________________________ # 3 _______________________________________ # 4 _______________________________________ # 5 _______________________________________ # 6 _______________________________________ # 7 _______________________________________ # 8 _______________________________________ # 9 _______________________________________ # 10 (least important) _______________________________________ .eli .page .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 This page intentionally left blank. .page .s 1 The following two questionnaires were distributed at DECUS by Kathie Norris of the VAX/VMS marketing group. Please fill them out carefully and return them to Kathy. These questionnaires are one of the best mechanisms we have to communicate our needs to the product planners. .page .s 2 .s 2 .s 2 .s 2 .s 2 .s 2 This page intentionally left blank. .page .c 70;QUESTIONNAIRE .c 70VAX/VMS SOURCES .s 1 To help us better understand your requirements relative to sources of the VAX/VMS please fill out the questionnaire below and return it within the next two weeks to: .s 1 .c 70;Kathie Norris (TW/A08) .c 70;Digital Equipment Corporation .c 70;1925 Andover Street .c 70;Tewksbury, Mass. 01876 .s 1 Name__________________________________________________ Title__________________________________________________ .s 1 Company__________________________________ Company Address________________________________________ .s 1 ##################################____________________________________________________ .br My primary application area is: .literal ____ Education ____ Resale ____ Engineering ____ Industrial ____ Government ____ Medical ____ Laboratory ____ Services ____ Telephone/Utilities ____ Other .end literal Please describe VAX-11/780 System Configuration installed or on order and what it is (or will be) used for. .s 1 .s 2 1. Do you have or do you plan to order a source kit for VAX/VMS? .s 1 #######No__________ #####Yes__________#### (If no, you don't need to fill out rest of questionnaire but please return) .s 1 2. In what form do you require the sources? .s 1 ###Machine readable###__________#####Listing only#######__________ .s 1 3. If you currently have a VAX/VMS source kit: .s 1 a. Has it met your expectations? Yes__________ No__________ .s 1 b. If not what additional requirements do you have? .s 1 Describe________________________________________________________________________________________ .s 1 4. What priority would you place the inclusion of sources for the following VAX/VMS components? (Use 1 as highest priority) .s 1 #####VAX/VMS Nucleus##__________ .br #####VAX/VMS Drivers##__________ .br #####Run Time Library#__________ .br #####Utilities########__________ .br #####RMS##############__________ .br #####Other############__________ .page 5. For what purpose do you require the sources? .s 1 Company policy#####__________ ###Backup#############__________ .s 1 Rebuild operating system with substitute modules###__________ ########(Minor ______ or ______ Major change) .s 1 To understand interfacing additional modules#__________ .s 1 Use to learn how system works__________ .s 1 Use as model for other code e.g. drivers__________ .s 1 Use to include minor changes to existing module__________ .s 1 Other_______________________________________________________ .s 1 6. How important is the VAX/VMS Source Kit? Please indicate price range you would be willing to pay. .s 1 ######################Machine#Readable#######Listing# .s 1 #####Very important###____________________________########____________________________ .br ###########Required###____________________________########____________________________ .br #######Nice to Have###____________________________########____________________________ .s 1 7. What documentation do you require for machine readable source kits? .s 2 .s 2 8. Do you need to have the source kits updated? .s 1 ##############Yes__________###No__________ .s 1 8a) How frequently? Each major release of VAX/VMS ______________ .s 1 ################Annually ____________ Quarterly ______________ .s 1 8b) How much would you pay for these update services? .s 1 9. Other comments _______________________________________________________________________________ .s 2 Thank you for your cooperation in completing the above questionnaire. .page .c 70;QUESTIONNAIRE .c 70;VAX/VMS BACKUP/RESTORE UTILITY .s 1 To help us better understand your detailed requirements relative to backup and restore under VAX/VMS please fill out the questionnaire below and return it within the next two weeks to: .s 1 .c 70;Kathie Norris (TW/A08) .c 70;Digital Equipment Corporation .c 70;1925 Andover Street .c 70;Tewksbury, Mass. 01876 .s 1 Name__________________________________________________ Title__________________________________________________ .s 1 Company__________________________________ Company Address________________________________________ .s 1 ##################################____________________________________________________ .br My primary application area is: .s 1 .literal ____ Education ____ Resale ____ Engineering ____ Industrial ____ Government ____ Laboratory ____ Services ____ Medical ____ Telephone/Utilities ____ Other .end literal Please describe VAX-11/780 System Configuration installed or on order and what it is (or will be) used for. .s 2 .s 2 1) Is DSC, volume backup/restore, adequate for your needs? .s 1 #####################Yes__________######No__________ .s 1 2) If not, please describe what area the product is deficient. .s 1 .lm 19 .i -10 ____________: Performance: What performance goals are necessary? (describe metrics: i.e. mbytes/minute) .s 1 What level of functionality is necessary? (e.g. incremental/file backup) .s 1 .i -10 ____________: Flexibility a) What devices are required to be backed up for backup device: .s 1 .s 2 b) What selection criteria for saving/restoring files is required? .s 2 .s 2 .i -10 ____________: Other .s 2 .lm 9 .s 1 3) How frequently do you backup the following: .br ###########(daily(D), weekly(W), monthly(M))? .s 1 a) System disk________ b) Data Files________ c) Program Files________ .s 1 4) When you back up data or program files, what is the amount of information backed up at one time? (All of volume, X% of volume, # files, etc). .s 2 5) How much time are you willing to spend backing up your data? (mb/min per backup type: system disk, volume, part of volume, selected files) .s 2 6) Must backup occur on line? .s 1 ############################Yes__________######No__________# .br Please#describe application use? .s 2 .s 2 7) What resources are you willing to dedicate per backup type: system disk, volume, part of volume, selected file(s)? (Amount of memory, _# of drives) .s 2 8) What ratio of capacity of data/system device to capacity of backup device is required? (_# media for backed up data) .s 2 9) What are the most frequently used data/backup device combinations? (% frequency) .s 1 disk/disk ________ disk/tape ________ disk/floppy disk ________ .s 1 10) What is the size of your data base (_# spindles, MB capacity)? .s 1 a) Do you expect it to grow over time? Yes__________ No__________ .s 1 b) If yes, estimate % growth per year ________________ .s 1 11) If a performance oriented incremental backup/restore capability were available, would current DSC be adequate for volume backup functionality? .s 1 #####################Yes__________ No__________ .s 1 12) If not, why not? .s 2 .s 2 13) What system has satisfactory Backup/Restore? Describe .s 1 #########DEC#supplied########Other vendor supplied .s 2 .s 2 14) a) In the event of a disk hardware failure that causes the media to be unreadable, would you purchase DEC services to attempt to recapture the data at the factory? .s 1 b) Would your requirement for data recapture differ if media were fixed rather than removable? .s 1 c) How much would you be willing to pay for these services? .s 2 Thank you for your cooperation in completing this questionnaire.