hicks@WALKER-EMH.ARPA (Gregory Hicks COMFLEACTS) (10/01/87)
Info-IBMPC Digest Wed, 30 September 1987 Volume 6 : Issue 66 This Week's Editor: Gregory Hicks - Chinhae Korea <hicks@walker-emh.arpa> Today's Topics: OS/2 Developers Kit Mixed language programs problem in linking MS Windows and IBM Mouse on PS/50 (2 msgs) PKARC Serious Squash Bug NDMAKE 4.3 - make utility for MSDOS now available HD Park Software available from SIMTEL20 Commercial Programs Redirecting LPT# to disk PRTEGA.ASM Updated Unix compatible tools source and binarys from SIMTEL20 CapsCtrl & Buf160 Utilities for new keyboards Today's Queries: Crossword Puzzles Eagle-PC info needed Reformat 1.50 trashing logical disks under DOS 3.3 FORTRAN 'interval arithmetic' compiler query Modifying a Program to read Drive C: Adding a 1.2 MB 5.25" floppy to an IBM PC/XT Adding a 3.5 inch drive to AT&T 6300 Sunshine EPROM Burner Software needed 80386 motherboard upgrade to IBM-PC/xt Adding Memory to an IBM EGA Card Experience with RightWriter/Rightwords Uncompress for unix .Z wanted MultiSync monitor with Hercules INFO-IBMPC BBS Phone Numbers: (213) 827-2635 (213) 827-2515 --------------------------------------------------------------------- Date: Sat, 26 Sep 87 14:10:54 PDT From: bmarsh@cod.nosc.mil (William C. Marsh) Subject: OS/2 Developers Kit >From: Anonymous >I've used OS/2 (Version 10.0, Rev 3.42.1.1) for approximately two months on both an IBM PC-AT model 139 (boosted to 8 MHz, 1 ws; 2.5 Mb ram; 32 Mb, 46 ms, I2L disk), and an AT clone (8 MHz, 0 ws; 2 Mb ram; 72 Mb, 24 ms, Toshiba disk. I've given up on OS/2. Why? My setups: AST Premium AT clone (10 Mhz, 0 ws), AT clone (10 Mhz, 1 ws). Both with approx. 2 Meg of ram. >1. OS/2 is slooooow. For example, OS/2 scrolls text on a MDA in real mode at about 2400 baud. I am using the same version (3.42.1.1 87/05/08), and I don't see this 'slowness'. I am using a Hercules compatable MDA, and text 'screams' by. The version of jove I have ported to OS/2 talks through DOS (as in write(1...) and I have no intention to change to using the video services provided by OS/2, it's fast enough as it is. >2. OS/2 is buggy. I experienced lots of mysterious crashes with lots of mysterious messages, i.e., "System detected an internal processing error", "Internal thread attempting to leave kernel", etc. OS/2 would even crash (for no reason at all) in the middle of boot. Mostly the 'thread attempting to leave kernel' error is caused by linking to the wrong c library (model size), then trying to run codeview to see why it doesn't work. I haven't had that error at any other time. But that's not to say there are no bugs: Trying to do anything (copy, dir, etc.) with either of my PC/VI (V 1.11) distribution disks causes an inter- nal processing error... >3. OS/2 is huge. Agreed. I'll wait to pass any judgements until the real distribution oc- curs. > And what do you get for 7 Mb of disk space? Not much. Only the bare basics, e.g., CD, COMP, COPY, DEL, DIR, EDLIN, MORE, REN, etc., are in- cluded with OS/2 (MORE still requires a "<"). Plus two C compilers, CodeView, MASM, sdked, bind, all the libraries... > "MS OS/2 will run ... with the following hardware: > ... > 1 meg of memory (2-3 recommended). > ..." > > Don't believe it. Don't even think about trying to use OS/2 with less > than 2 Mb ram. The C compiler, linker, etc. won't run in less. I ran fine with 1 meg in my AST until I could put more in. >4. OS/2 won't boot with foreign cards installed, such as a 3Com Ethernet > card... I have a 3Com Ethernet card in my AST, and it boots fine. Sounds to me that you have the card sharing an interrupt or DMA channel. I also have an ARNET 8 port serial card installed in the same computer. Haven't had to switch any boards, except the AST EGA card (which just needs a new ROM, as we learned from the AST guy at the conference). >5. Many of OS/2's "new" features did not work. At least Microsoft was kind enough to warn you of some, > "...memory swapping to disk is not solid in this release. If you > stress it too hard it will fail and the system will stop." > Hey, I couldn't get swapping to work AT ALL. Codeview doesn't under- stand the concept of threads, nor can it debug dynamic link libraries. The list goes on and on. And that's what MS was nice enough to acknowledge--- LOTS of OS/2 just doesn't work. CodeView needs work, as it can crash the system. But other than that, I haven't had any problems, associated with swapping (which I use) or others. I have tries queues, pipes (jove uses them for running other processes), and threads with no problems. The C library needs to support multi-thread programs better, though... >6. OS/2 uses the same pathetic file system found in MS-DOS... Do you want to re-format your hard disk everytime you switch operating systems? If I need to run DOS 3.3 (to use our network or tape backup) I simply boot from a diskette. Make life very easy for me. If you want UNIX 4.2 BSD, then I would suggest talking to Berkeley. >Conclusion >OS/2 was a terrible waste of $3000 and 2 months of my time. I'm sorry you feel that way. It has been a very enjoyable three weeks for me, and I'm looking forward for more. I really think it is childish to think Microsoft is trying to rip off 2000+ people. If the developers all thought like you, they wouldn't be able to sell the $395 operating system to anybody. >And that's just for the beta version---when the real OS/2 comes out, >you get them to pay you again! About $395. Developers will get all releases, up to the OS/2 version 1.1 release. Be- sides, do you remember paying $200 for CP/M? I do. >But wait, there's more! You'll need new OS/2 software too. The C compiler, MASM, and sdked come with the developers kit now. All the others work in the dos compatability box anyway, so why do you feel you 'have' to upgrade? >P.S., if you own a non-IBM keyboard destined for OS/2 use, you might want to test it to make sure that it handles ctl-esc and alt-esc, the key com- binations used to bring up the session manager and switch between screen groups. Alt-esc doesn't work on many AT-clone keyboards! Certainly not Microsoft's fault that your clone keyboard doesn't work. All my clone keyboards work... Bill Marsh, Naval Ocean Systems Center, San Diego, CA {arpa,mil}net: bmarsh@cod.nosc.mil uucp: {ihnp4,akgua,decvax,dcdwest,ucbvax}!sdcsvax!nosc!bmarsh "If everything seems to be coming your way, you're probably in the wrong lane." ------------------------------ Date: Sat, 26 Sep 1987 21:40-EDT From: kwh@di.sei.cmu.edu.UUCP (Kurt Hoyt) Subject: Mixed language programs problem in linking Organization: Carnegie-Mellon University, SEI, Pgh, Pa In article <582@uvm-gen.UUCP> haviland@uvm-gen.UUCP (Tom Haviland) writes: > > I have an interesting problem I wonder if anyone out there might have >experienced. I'm writing a menu system to go with a fortran modelling >package, and since supposedly microsoft C and FORTRAN work together, I >wrote the menu system in C. I've set up an INTERFACE block in the fortran >program to call the C routines, and it seems to call the routine ok, but >when the C routine calls the fscanf() function, reading floating point >numbers from a file (%f) I get an error "Floating point not loaded" and >the program stops. Now the C routines do some floating point calculations > Thomas P. Haviland Your problem isn't in the compiling -- its in the linking. If you link from FORTRAN instead of from cc, the proper library for scanf/printf for float- ing point numbers isn't linked. See page 367 of the C compiler user's guide manual: "Your program needs the foating-point library, but the library was not loaded. The error causes the program to terminate with an exit status of 255." You need to link in the C floating point library when you link. FORTRAN won't know that you need the C floating point library. Hope this helps. It is a common error when mixing languages. I ran into similar problems mixing C and Ada because Ada didn't link in some startup code that the C code needed, so I had to explicitly tell Ada to link in the C startup code. ------------------------------ Date: Sun, 27 Sep 87 00:52:29 PDT From: Jeff Wallace <jeffw%con.Berkeley.EDU@ucbvax.Berkeley.EDU> Subject: MS Windows and the IBM Mouse on PS/50 Cc: Doug_Wade%UBC.MAILNET@mit-multics.arpa >I've had no luck trying to get Microsoft Windows to recognize a IBM mouse >on my PS/50. I had the same problem with my model 60 when I purchased it in May. The problem I had was that I had Windows v1.03. Version 1.03 does not have the proper drivers for the PS/2 mouse and the 8512, 8513, 8514 and 8514/A monitors. You can substitute one of the video drivers (I think the EGA one) and it will work fine, but the mouse is a different story. I called Aldus Corp. (the makers of PageMaker) and found out from technical support that they had created a version 1.0a of PageMaker that incorporated the new Windows 1.04 drivers. Great, I thought. But no. Aldus no longer distributes the full Windows package, only the run-time package. I got that for $10 (plus one Federal Express charge) in two days. I then took the new drivers and various other new files from the Aldus stuff and mixed it with my 1.03 files, shook vigorously and prayed. Finally, I got my PS/2 mouse working with Windows and the 1.04 bug fixes applied to Windows (some PS/2 specific problems apparently). I called Microsoft and found out that there exists a version 1.04 which you can get for free if you are a registered owner of 1.03. I later found out that the drivers are available on Compuserve in Microsoft's forum. That didn't help much since I don't subscribe to Com- puserve. Jeff Wallace ARPA/Internet: jeffw@scam.Berkeley.Edu U.C. Berkeley UUCP: ...!ucbvax!scam!jeffw ------------------------------ Date: 26 Sep 87 16:34 +0200 From: "Niklas Mollberg" <P15738%QZCOM.BITNET@wiscvm.wisc.edu> Subject: MS-Mouse and Windows on PS/2 Answer to: Doug_Wade%UBC.MAILNET@MIT-Multics.ARPA IF you have a Windows version 1.00, 1.01, 1.02 or 1.03 you must get a copy of Windows 1.04 or the MOUSE.DRv from that version. Windows does not recognize any other drivers (MOUSE.COM, MOUSE.SYS) - Niklas ------------------------------ Date: Sunday, 20 September 1987 02:41-MDT From: Phil Katz <katz@UWM-CS.MILW.WISC.EDU> Subject: PKARC Serious Squash Bug I have been able to confirm the bug in PKXARC 3.5 if a file (uncompressed) starts with 0x90, and the file is Squashed. To date, I have received two direct reports of this problem. For trivia buffs, this bug was also present in PKXARC 3.4, and never reported. If a file starts with a 0x90 character (144 decimal) and is Squashed, PKXARC (and PKXARCjr and PKSFX) will not output this character, and will report the file fails CRC check. The resulting extracted file will be ex- actly 1 byte shorter than it should be (it is missing the 0x90 character at the beginning). Except for missing the first character, the extracted file will compare exactly with the original. Re-inserting a 0x90 character at the beginning of the extracted file will restore it to the original file. This problem can be fixed by the applying the following patches to the PKXARC/PKXARCjr/PKSFX version 3.5 programs: debug pkxarc.com e 1d0b 8b 3e c8 f4 80 3e d0 f5 0c 75 06 e8 a9 06 eb 1a 90 aa w q debug pkxarcjr.com e 1c92 8b 3e aa f5 80 3e b2 f6 0c 75 06 e8 a8 06 eb 1a 90 aa w q debug pksfx.pgm e 112 8c f5 e 16a4 8b 3e 2a f7 80 3e 24 f8 0c 75 06 e8 17 04 eb 1a 90 aa w q >Phil Katz> Exec-PC BBS, Shorewood WI ......... 414-964-5160, 24 hours, 1200/2400 baud RBBS of Chicago ................... 312-352-1035, 24 hours, 1200/2400 baud Sound of Music BBS, Oceanside NY .. 516-536-8723, 24 hours, 1200/2400 baud PKWARE BBS ........................ 414-352-7176, 24 hours, 1200/2400 baud USENET ............................ katz@uwm-cs.milw.wisc.edu BIX ............................... Username: philkatz U.S. Mail ......................... 7032 N. Ardara Ave., Glendale, WI 53209 ------------------------------ Date: Monday, 21 September 1987 23:41-MDT From: kneller@CGL.UCSF.EDU (Don Kneller) Subject: NDMAKE 4.3 - make utility for MSDOS now available Now available from SIMTEL20... Filename Type Bytes CRC Directory PD:<MSDOS.C> NDMAKE43.ARC.1 BINARY 54214 30B6H NDMAKE (ver. 4.3) is a shareware make utility for MSDOS. For those un- familiar with `make', it is a utility that automates development of other programs. This MSDOS version includes: - auto response file generation to overcome the command line limit - VPATH (find sources in alternate directories) - full macro capability - default rules - looping over commands Don Kneller UUCP: ...ucbvax!ucsfcgl!kneller INTERNET: kneller@cgl.ucsf.edu BITNET: kneller@ucsfcgl.BITNET ------------------------------ Date: Sun, 27 Sep 87 15:50:10 EDT From: David Kirschbaum <kirsch@braggvax.arpa> Subject: HD Park software available from SIMTEL20 NetLandians, Recently read a query about PD disk park utilities on this newsletter. I'd suggest snarfing (all from SIMTEL20): PD:<MSDOS.DSKUTL> PDTIMPRK.ARC 9 kb TIMEPARK.ARC 4 kb TIMPARK1.LBR 4 kb and maybe?? DISKP.EXE 8 kb I snarfed one of these a while ago, works just fine, but can't remember which one. Haven't looked at the contents lately to see what's there, but expect nice assembler source (right?). Coulda sworn there was one clean utility to park on command, but can't find it. David Kirschbaum Toad Hall kirsch@braggvax.ARPA ------------------------------ Date: 28 Sep 87 11:35:23 GMT From: hsi!tankus@uunet.UU.NET (Ed Tankus) Subject: Commercial Programs Redirecting LPT# to disk There is a commercial product called PrintQ that retails for $89. I think it does exactly what you are looking for. It allows any print program to be spooled to disk for printout at a later date. The product has received favorable reviews from PC MAG and others. The company is in NJ. I don't have a phone number but they advertise in most PC magazines anyway. Cheers! -- Ed. Net : {uunet,ihnp4,noao,hao,yale}!hsi!tankus Snail: Health Systems Int'l, 100 Broadway, New Haven, CT 06511 Bell : (203) 562-2101 [The program LPTX.ASM in the Info-IBMPC Lending Library also works, is free, and is source code. --gph] ------------------------------ Date: Fri, 25 Sep 87 15:55 CET From: Helmut Waelder <ZRWA001%DTUZDV1%DEARN.BITNET@wiscvm.wisc.edu> Subject: PRTEGA.ASM I have added code to reset the printer to 6 lpi after graphics output. The complete source file follows: [PRTEGA.asm has been updated in the lending library - gph] ------------------------------ Date: Wednesday, 23 September 1987 07:17-MDT From: mcvax!dutrun!dutesta!karel@SEISMO.CSS.GOV (Karel van Houten) Subject: Unix compatible tools source and binarys from SIMTEL20 Now available from SIMTEL20... Filename Type Bytes CRC Directory PD:<MSDOS.DIRUTL> TOOLS.ARC.1 BINARY 69883 3B14H Some time ago there was a request for a UNIX compatible 'mv' command for MS-DOS. As a UNIX addict, I have written some commonly used UNIX tools in Microsoft C / MASM for myself. Here are the executables and sources for those who want to try them too. This package consist of: df, du, ls, mv, rm. -- Karel van Houten BITNET : NO! NO! NO! Delft University of Technology INTERNET : karel@dutesta.UUCP Faculty of Electrical Engineering UUCP : ..!mcvax!dutrun!dutesta!karel Room 9.29H, Mekelweg 4 SURFnet : DUTRUN::KAREL 2628 CD DELFT, The Netherlands VOICE : +31 15 783502 ------------------------------ Date: Mon, 28 Sep 87 21:23 EDT From: Hess.Unicorn@MIT-Multics.ARPA Subject: CapsCtrl & Buf160 Utilities, for new keyboards Mailed under separate cover for the Info-IBMPC library are two pieces of source code, both intended for effective use of the new almost-not- braindamaged IBM keyboards. The first, CAPSCTRL.ASM, makes the Caps Lock key turn into a control key. To get a "real" Caps Lock press, do Shift-CapsLock or Alt-CapsLock. [CAPSCTRL.ASM got lost in the Mail System Somewhere. gph] The second, BUF160.ASM, is a replacement for the existing one. It works with the new keyboards, and enables the use of the F11 and F12 keys as well as the usual 160 character buffer. No ill effects from using this with an old keyboard. Both CapsCtrl and the Buf160 changes are courtesy of Bill Spitzak at Mark of the Unicorn, Inc. Brian [BUF160NEW.ASM has been added to the Info-IBMPC Lending Library - gph] ------------------------------ Date: Sat, 26 Sep 87 09:26:57 EDT From: David Kirschbaum <kirsch@braggvax.arpa> Subject: Crossword Puzzles NetLandians, A critic recently asked me, "Well, if your computer is so smart, how come it can't do crossword puzzles?" Got into a discussion, suddenly realized I'd NEVER seen any code around to have a micro MAKE a crossword puzzle! Shouldn't be TOO hard .. especially in one of the recursive-friendly lan- guages like LISP or something, I'd think .. can anyone point me to some public domain source (any language, I guess) that creates crossword puzzles from word lists or dictionaries? Thanks in advance, David Kirschbaum Toad Hall kirsch@braggvax.ARPA [A cross word puzzle generator (written in FORTRAN) appeared in an issue of "Dr. Dobb's Journal" in the late 70's. I have the program somewhere in one of the boxes from my numerous moves. If I can find it, I'll send it to you and the Lending Library. I also saw a program (also written in FORTRAN) in a somewhat later issue of Dr. Dobb's that generated hidden word puzzles but for the life of me, I don't know where it went to... Does anyone have back issues of Dr. Dobb's? gph] ------------------------------ Date: Sat, 26 Sep 87 23:27:50 EDT From: weikert@nadc.arpa (J. Weikert) Subject: Eagle-PC info needed I know someone who has acquired an old Eagle-PC computer and would like to put a standard 20 meg. hard disk system in it. He has not been successful so far. Does anyone know anything about these machines? ------------------------------ Date: Sun, 27 Sep 87 10:29 N From: <IUS%DACTH51.BITNET@CUVMA.COLUMBIA.EDU> Subject: Reformat 1.50 trashing logical disks under DOS 3.3 I have been using your excellent Reformat program for about one year or so without any major problems. (I do backup, but so far, this has not proven necessary) One thing though. I have found that REFORMAT runs quite well under DOS 3.3 without any glitchs. However, if I want to use that new feature of 3.3 FDISK to have two logical disks on one physical (let's say a 33MB+) disk and reformat either of those logical disks then it trashes the other (logical) disk. Any ideas? Fixes? regards, el Eberhard W. Lisse Department of Plastic and Reconstructive Surgery and Burn Unit Technical University Aachen, West Germany ------------------------------ Date: Mon, 28 Sep 87 21:51:11 CDT From: "Rich Winkel UMC Math Department" <MATHPG1%UMCVMB.BITNET@wiscvm.wisc.edu> Subject: FORTRAN 'interval arithmetic' compiler query Does anyone know of such a beast for the PC? For the record, such a com- piler automatically keeps track of computational error bounds as the errors propagate through arithmetic calculations, so that results can be derived in the form 'answer= 1.22323 +- .00002'. Thanks for any help! Rich Winkel (MATHPG1@UMCVMB.BITNET) ------------------------------ Date: Thu, 24 Sep 87 22:55:40 MDT From: dd@lanl.gov (Dan Davison) Subject: Program always reads from drive A: I have an application program, available only as a .EXE file, which always tries to read a data file from the A: drive. Is there a way to change this using DEBUG? I'd like to change this to the C: drive. This program is not a commerical program, and source code is not available so that I can fix this problem. Any hints would be welcome. Dan Davison Arpa: dd@lanl.gov BITNET: dd%lanl.gov@wiscvm.bitnet CIS: 74065,41 UUCP: dd@lanl.uucp or ...cmcl2!lanl!dd Voice:505-665-1355| T-10 MS K710, Los Alamos Natl. Lab, Los Alamos,NM 87545 ------------------------------ Date: Thu, 24 Sep 87 22:55:58 MDT From: dd@lanl.gov (Dan Davison) Subject: How to add a 1.2 MB 5.25" floppy to an IBM PC/XT I have asked a local computer store if it is possible to add a 1.2 MB floppy drive to an IBM PC/XT (actually, it's an IBM PC Portable, which has the XT motherboard). After about a month they tell me this is not pos- sible. I'm aware that I'll have to get a different controller card, but surely there is a 1.2MB floppy controller for the PC bus. Does anyone have a high density floppy in an XT, and if so can you recommend a controller? Dan Davison Arpa: dd@lanl.gov BITNET: dd%lanl.gov@wiscvm.bitnet CIS: 74065,41 UUCP: dd@lanl.uucp or ...cmcl2!lanl!dd Voice:505-665-1355| T-10 MS K710, Los Alamos Natl. Lab, Los Alamos,NM 87545 ------------------------------ Date: Thu, 24 Sep 87 16:57:57 EDT From: Glen Jones <gdj5t@krebs.acc.virginia.edu> Subject: Adding a 3.5 inch drive to AT&T 6300 Is there any way to add an external 3.5 inch drive to an AT&T 6300 PC ***AND*** keep both internal 5.25 floppy drives? If not, are there any cautions involving buying/installing an internal 3.5 in place of a 5.25 inch? I would really like to keep both 5.25 inch drives, however. I am currently running PC-DOS 3.2 (by IBM) on the 6300; has anyone done this with a 3.5 inch drive on an AT&T6300? Details on my current setup follow: ROB bios 1.43 640K RAM HardCard20 installed (from Plus Dev.) both serial and parallel ports used Thanks in advance for ANY advice! Glen Jones U.Va. Biochem Dept., Box 440 > UUCP: ...!uunet!virginia!gdj5t Charlottesville, VA 22908 > BITNET: gdj5t@virginia.BITNET (804) 924-2373 > CSNET: gdj5t@acc.virginia.edu ------------------------------ Date: Fri, 25 Sep 87 17:03 EST From: <JWS%PSUARLC.BITNET@wiscvm.wisc.edu> Subject: Sunshine EPROM Burner We purchased a Sunshine EPROM burner through third parties. The accompany- ing software does not function on an EGA, but works OK on CGA. Do you know an address or telephone number for this company or a "quick fix" for the problem? Thanks in advance, Jack Sharer, JWS@PSUARLC ------------------------------ Date: Tue, 29 Sep 87 23:11:42 EDT From: Dimitri Vulis <CLDLV%NASAGISS.BITNET@WISCVM.WISC.EDU> Subject: Measuring speed from Turbo Pascal-bug fixed > From: RECK%DBNUAMA1.BITNET@wiscvm.wisc.edu (Gisbert W.Selke) > Subject: Measuring speed from within Turbo Pascal > > I need to measure the approximate clock rate of a PC from within a Turbo > Pascal program. So I wrote a little procedure which first got the cur- rent > time (via a DOS call), then executed some silly little loop and then got > the time again. > > >>> First question: Is there a simpler or more intelligent way to find the > speed? Consider the following: program measurespeed; function speed:byte; const NN=3; begin inline( $1E (* push ds *) /$31/$C0 (* xor ax,ax *) /$8E/$D8 (* mov ds,ax *) /$8A/$1E/>$46C (* mov bl,46c *) /$88/$DF (* mov bh,bl *) /$80/$C7/<NN (* add bh,NN *) /$3A/$1E/>$46C (*L1:cmp bl,46c *) /$74/$FA (*jz L1 *) /$40 (*L2:inc ax *) /$3A/$3E/>$46C (*cmp bh,46c *) /$75/$F9 (*jnz L2 *) /$1F (*pop ds *) /$88/$66/$04 (*mov bp+4,ah *) ) end; begin writeln('Speed returns ',speed); end. ALGORITHM: program measurespeed; const NN=3; var z:byte; count:integer; speed:byte; begin z:=mem0:$46C; (* or use 'absolute'... *) while z=mem0:$46C do; (* here we start a new tick *) z:=(mem0:$46C+NN) mod 256; (* wait for this value *) count:=0; repeat count:=count+1; until z=mem0:$46C; speed:=count div 256; end. EXPLANATION: On every PC-compatible, the timer interrupt (IRQ0) occurs 18.2 times per second and the handler routine increments the 4-byte value at 0:46Ch. We get its least significant byte, wait until it changes, and then increment a counter until the timer reaches t+NN, where NN is a number like 3 or 5. We return the MSB of the counter (i.e. counter mod 256) only, since the LSB is sometimes inconsistent (if a keyboard interrupt etc occurs, it steals some CPU time I guess). ADVANTAGES OVER PASCAL CODE: 1) The counter has to be 16-bit. When TP allocates it, it's as likely as not to fall on odd address, with significant performance difference on a 16-bit machine, not every time you run it, but when you change the program someplace else, so the address changes between even and odd. (the only real advantage) 2) The instruction mix generated for the repeat-until is a lot richer than loop L2, making the number of cycles more CPU model dependent. 3) On a very fast machine, the counter may overflow. We treat AX as un- signed. RELIABILITY: This function seems to give consistent results every time. This result is proportional (with a large error margin) to some sort of performance coef- ficient. Obviously, the result is affected by prefetch queues, wait states etc a whole lot. It's not really the clock speed, but some number propr- tional to some sort of 'performance index'. It's about as (un)reliable as Norton SI. ------------------------------ Date: Sun, 27 Sep 87 17:12:26 PDT From: Ya'akov_Miles%UBC.MAILNET@MIT-Multics.ARPA Subject: 80386 motherboard upgrade to IBM-PC/xt On Page 36, of "Computing Now!", Sept. 1987, is advertised an 80386 mother- board that is a drop-in replacement for the standard IBM-PC/xt motherboard. This 80386 motherboard runs at 6/16 mHz with zero wait states, and features four 16 bit and four 8 bit slots. Made by: Chie Dar Electronic Ind. Co., Ltd. Telex 34181 P.O. Box 89-36, Taipei, Taiwan, ROC Phone (02) 981-4487 Does anybody have any further information - it may be a cheap and dirty way to upgrade a plain vanilla XT-clone. A super "BABY-AT"? ------------------------------ Date: Tue, 29 Sep 87 13:06:45 edt From: moss!cbosgd!neoucom!wtm@RUTGERS.EDU (Bill Mayhew) Subject: Adding Memory to an IBM EGA Card One of our faculty people would like to increase the memory on his real IBM brand EGA card from 128K to 256K. He has the daughter board for holding the additonal chips. Alas, the board was installed by a dealer who was not kind enough to send along the installation notes. It looks as if the chips are just plain old 4164, 150 nS chips, but I want to be sure before I take screwdriver in hand (the IBM-supplied chips have some unrecognizable in-house part no.). When I add the extra memory, are there any jumpers that need to be changed? Thanks much to anybody who can assist. --Bill Bill Mayhew, Electrical Engineering Division of Basic Medical Sciences Northeastern Ohio Universities' College of Medicine Rootstown (what a name!), OH 44272-9989 USA phone: 216-325-2511 (wtm@neoucom.UUCP ...!cbosgd!neoucom!wtm) ------------------------------ Date: Tue, 29 Sep 87 16:51:00 MEZ From: Z3000JD%AWITUW01.BITNET@wiscvm.wisc.edu (Johannes Demel Digitalrechenzentrum) Subject: Experience with RightWriter/Rightwords Does anyone has experience with the Grammer and Style-Checker RightWriter and Dictionary-Utility RightWords from RIGHTSOFT Inc. (At the moment of- fered by IEEE at a reduced price). I am most interested using this in an application with the TeX and LaTeX Typesetting system by Donald Knuth. Johannes Demel, Technical University Vienna (BITNET: Z3000JD@AWITUW01) ------------------------------ Date: Tue 29 Sep 87 14:53:04-PDT From: Ted Shapin <BEC.SHAPIN@ECLA.USC.EDU> Subject: Uncompress for unix .Z wanted Phone: (714)961-3393; Mail:Beckman Instruments, Inc. Mail-addr: 2500 Harbor Blvd., X-11, Fullerton CA 92634 I would like to uncompress some unix .Z files. Is there such a program that runs under PC-DOS? ------------------------------ Date: Wed, 30 Sep 87 14:25:44 MEZ From: "Rainer Kleinrensing" UNI217%DBNRHRZ1.BITNET@WISCVM.WISC.EDU Subject: MultiSync monitor with Hercules Some time ago I bought a QuadEGA ProSync graphics card and a Multisync monitor from Quadram. I have read several times that Multisync compatible monitors can operate with a Hercules monochrome graphics card. The ProSync card is capable of emulating a Hercules card. My question: how do I have to connect the monitor to the card in order to operate in Hercules mode? The manuals don't give any clues. I have heard that all I have to change is the monitor cable. If anybody out there successfully operates a Her- cules card with a MultiSync monitor, please let me know. Thank you , Rainer ------------------------------ End of Info-IBMPC Digest ************************
hicks@WALKER-EMH.ARPA (Gregory Hicks COMFLEACTS) (10/01/87)
Info-IBMPC Digest Wed, 30 September 1987 Volume 6 : Issue 66 This Week's Editor: Gregory Hicks - Chinhae Korea <hicks@walker-emh.arpa> Today's Topics: OS/2 Developers Kit Mixed language programs problem in linking MS Windows and IBM Mouse on PS/50 (2 msgs) PKARC Serious Squash Bug NDMAKE 4.3 - make utility for MSDOS now available HD Park Software available from SIMTEL20 Commercial Programs Redirecting LPT# to disk PRTEGA.ASM Updated Unix compatible tools source and binarys from SIMTEL20 CapsCtrl & Buf160 Utilities for new keyboards Today's Queries: Crossword Puzzles Eagle-PC info needed Reformat 1.50 trashing logical disks under DOS 3.3 FORTRAN 'interval arithmetic' compiler query Modifying a Program to read Drive C: Adding a 1.2 MB 5.25" floppy to an IBM PC/XT Adding a 3.5 inch drive to AT&T 6300 Sunshine EPROM Burner Software needed 80386 motherboard upgrade to IBM-PC/xt Adding Memory to an IBM EGA Card Experience with RightWriter/Rightwords Uncompress for unix .Z wanted MultiSync monitor with Hercules INFO-IBMPC BBS Phone Numbers: (213) 827-2635 (213) 827-2515 --------------------------------------------------------------------- Date: Sat, 26 Sep 87 14:10:54 PDT From: bmarsh@cod.nosc.mil (William C. Marsh) Subject: OS/2 Developers Kit >From: Anonymous >I've used OS/2 (Version 10.0, Rev 3.42.1.1) for approximately two months on both an IBM PC-AT model 139 (boosted to 8 MHz, 1 ws; 2.5 Mb ram; 32 Mb, 46 ms, I2L disk), and an AT clone (8 MHz, 0 ws; 2 Mb ram; 72 Mb, 24 ms, Toshiba disk. I've given up on OS/2. Why? My setups: AST Premium AT clone (10 Mhz, 0 ws), AT clone (10 Mhz, 1 ws). Both with approx. 2 Meg of ram. >1. OS/2 is slooooow. For example, OS/2 scrolls text on a MDA in real mode at about 2400 baud. I am using the same version (3.42.1.1 87/05/08), and I don't see this 'slowness'. I am using a Hercules compatable MDA, and text 'screams' by. The version of jove I have ported to OS/2 talks through DOS (as in write(1...) and I have no intention to change to using the video services provided by OS/2, it's fast enough as it is. >2. OS/2 is buggy. I experienced lots of mysterious crashes with lots of mysterious messages, i.e., "System detected an internal processing error", "Internal thread attempting to leave kernel", etc. OS/2 would even crash (for no reason at all) in the middle of boot. Mostly the 'thread attempting to leave kernel' error is caused by linking to the wrong c library (model size), then trying to run codeview to see why it doesn't work. I haven't had that error at any other time. But that's not to say there are no bugs: Trying to do anything (copy, dir, etc.) with either of my PC/VI (V 1.11) distribution disks causes an inter- nal processing error... >3. OS/2 is huge. Agreed. I'll wait to pass any judgements until the real distribution oc- curs. > And what do you get for 7 Mb of disk space? Not much. Only the bare basics, e.g., CD, COMP, COPY, DEL, DIR, EDLIN, MORE, REN, etc., are in- cluded with OS/2 (MORE still requires a "<"). Plus two C compilers, CodeView, MASM, sdked, bind, all the libraries... > "MS OS/2 will run ... with the following hardware: > ... > 1 meg of memory (2-3 recommended). > ..." > > Don't believe it. Don't even think about trying to use OS/2 with less > than 2 Mb ram. The C compiler, linker, etc. won't run in less. I ran fine with 1 meg in my AST until I could put more in. >4. OS/2 won't boot with foreign cards installed, such as a 3Com Ethernet > card... I have a 3Com Ethernet card in my AST, and it boots fine. Sounds to me that you have the card sharing an interrupt or DMA channel. I also have an ARNET 8 port serial card installed in the same computer. Haven't had to switch any boards, except the AST EGA card (which just needs a new ROM, as we learned from the AST guy at the conference). >5. Many of OS/2's "new" features did not work. At least Microsoft was kind enough to warn you of some, > "...memory swapping to disk is not solid in this release. If you > stress it too hard it will fail and the system will stop." > Hey, I couldn't get swapping to work AT ALL. Codeview doesn't under- stand the concept of threads, nor can it debug dynamic link libraries. The list goes on and on. And that's what MS was nice enough to acknowledge--- LOTS of OS/2 just doesn't work. CodeView needs work, as it can crash the system. But other than that, I haven't had any problems, associated with swapping (which I use) or others. I have tries queues, pipes (jove uses them for running other processes), and threads with no problems. The C library needs to support multi-thread programs better, though... >6. OS/2 uses the same pathetic file system found in MS-DOS... Do you want to re-format your hard disk everytime you switch operating systems? If I need to run DOS 3.3 (to use our network or tape backup) I simply boot from a diskette. Make life very easy for me. If you want UNIX 4.2 BSD, then I would suggest talking to Berkeley. >Conclusion >OS/2 was a terrible waste of $3000 and 2 months of my time. I'm sorry you feel that way. It has been a very enjoyable three weeks for me, and I'm looking forward for more. I really think it is childish to think Microsoft is trying to rip off 2000+ people. If the developers all thought like you, they wouldn't be able to sell the $395 operating system to anybody. >And that's just for the beta version---when the real OS/2 comes out, >you get them to pay you again! About $395. Developers will get all releases, up to the OS/2 version 1.1 release. Be- sides, do you remember paying $200 for CP/M? I do. >But wait, there's more! You'll need new OS/2 software too. The C compiler, MASM, and sdked come with the developers kit now. All the others work in the dos compatability box anyway, so why do you feel you 'have' to upgrade? >P.S., if you own a non-IBM keyboard destined for OS/2 use, you might want to test it to make sure that it handles ctl-esc and alt-esc, the key com- binations used to bring up the session manager and switch between screen groups. Alt-esc doesn't work on many AT-clone keyboards! Certainly not Microsoft's fault that your clone keyboard doesn't work. All my clone keyboards work... Bill Marsh, Naval Ocean Systems Center, San Diego, CA {arpa,mil}net: bmarsh@cod.nosc.mil uucp: {ihnp4,akgua,decvax,dcdwest,ucbvax}!sdcsvax!nosc!bmarsh "If everything seems to be coming your way, you're probably in the wrong lane." ------------------------------ Date: Sat, 26 Sep 1987 21:40-EDT From: kwh@di.sei.cmu.edu.UUCP (Kurt Hoyt) Subject: Mixed language programs problem in linking Organization: Carnegie-Mellon University, SEI, Pgh, Pa In article <582@uvm-gen.UUCP> haviland@uvm-gen.UUCP (Tom Haviland) writes: > > I have an interesting problem I wonder if anyone out there might have >experienced. I'm writing a menu system to go with a fortran modelling >package, and since supposedly microsoft C and FORTRAN work together, I >wrote the menu system in C. I've set up an INTERFACE block in the fortran >program to call the C routines, and it seems to call the routine ok, but >when the C routine calls the fscanf() function, reading floating point >numbers from a file (%f) I get an error "Floating point not loaded" and >the program stops. Now the C routines do some floating point calculations > Thomas P. Haviland Your problem isn't in the compiling -- its in the linking. If you link from FORTRAN instead of from cc, the proper library for scanf/printf for float- ing point numbers isn't linked. See page 367 of the C compiler user's guide manual: "Your program needs the foating-point library, but the library was not loaded. The error causes the program to terminate with an exit status of 255." You need to link in the C floating point library when you link. FORTRAN won't know that you need the C floating point library. Hope this helps. It is a common error when mixing languages. I ran into similar problems mixing C and Ada because Ada didn't link in some startup code that the C code needed, so I had to explicitly tell Ada to link in the C startup code. ------------------------------ Date: Sun, 27 Sep 87 00:52:29 PDT From: Jeff Wallace <jeffw%con.Berkeley.EDU@ucbvax.Berkeley.EDU> Subject: MS Windows and the IBM Mouse on PS/50 Cc: Doug_Wade%UBC.MAILNET@mit-multics.arpa >I've had no luck trying to get Microsoft Windows to recognize a IBM mouse >on my PS/50. I had the same problem with my model 60 when I purchased it in May. The problem I had was that I had Windows v1.03. Version 1.03 does not have the proper drivers for the PS/2 mouse and the 8512, 8513, 8514 and 8514/A monitors. You can substitute one of the video drivers (I think the EGA one) and it will work fine, but the mouse is a different story. I called Aldus Corp. (the makers of PageMaker) and found out from technical support that they had created a version 1.0a of PageMaker that incorporated the new Windows 1.04 drivers. Great, I thought. But no. Aldus no longer distributes the full Windows package, only the run-time package. I got that for $10 (plus one Federal Express charge) in two days. I then took the new drivers and various other new files from the Aldus stuff and mixed it with my 1.03 files, shook vigorously and prayed. Finally, I got my PS/2 mouse working with Windows and the 1.04 bug fixes applied to Windows (some PS/2 specific problems apparently). I called Microsoft and found out that there exists a version 1.04 which you can get for free if you are a registered owner of 1.03. I later found out that the drivers are available on Compuserve in Microsoft's forum. That didn't help much since I don't subscribe to Com- puserve. Jeff Wallace ARPA/Internet: jeffw@scam.Berkeley.Edu U.C. Berkeley UUCP: ...!ucbvax!scam!jeffw ------------------------------ Date: 26 Sep 87 16:34 +0200 From: "Niklas Mollberg" <P15738%QZCOM.BITNET@wiscvm.wisc.edu> Subject: MS-Mouse and Windows on PS/2 Answer to: Doug_Wade%UBC.MAILNET@MIT-Multics.ARPA IF you have a Windows version 1.00, 1.01, 1.02 or 1.03 you must get a copy of Windows 1.04 or the MOUSE.DRv from that version. Windows does not recognize any other drivers (MOUSE.COM, MOUSE.SYS) - Niklas ------------------------------ Date: Sunday, 20 September 1987 02:41-MDT From: Phil Katz <katz@UWM-CS.MILW.WISC.EDU> Subject: PKARC Serious Squash Bug I have been able to confirm the bug in PKXARC 3.5 if a file (uncompressed) starts with 0x90, and the file is Squashed. To date, I have received two direct reports of this problem. For trivia buffs, this bug was also present in PKXARC 3.4, and never reported. If a file starts with a 0x90 character (144 decimal) and is Squashed, PKXARC (and PKXARCjr and PKSFX) will not output this character, and will report the file fails CRC check. The resulting extracted file will be ex- actly 1 byte shorter than it should be (it is missing the 0x90 character at the beginning). Except for missing the first character, the extracted file will compare exactly with the original. Re-inserting a 0x90 character at the beginning of the extracted file will restore it to the original file. This problem can be fixed by the applying the following patches to the PKXARC/PKXARCjr/PKSFX version 3.5 programs: debug pkxarc.com e 1d0b 8b 3e c8 f4 80 3e d0 f5 0c 75 06 e8 a9 06 eb 1a 90 aa w q debug pkxarcjr.com e 1c92 8b 3e aa f5 80 3e b2 f6 0c 75 06 e8 a8 06 eb 1a 90 aa w q debug pksfx.pgm e 112 8c f5 e 16a4 8b 3e 2a f7 80 3e 24 f8 0c 75 06 e8 17 04 eb 1a 90 aa w q >Phil Katz> Exec-PC BBS, Shorewood WI ......... 414-964-5160, 24 hours, 1200/2400 baud RBBS of Chicago ................... 312-352-1035, 24 hours, 1200/2400 baud Sound of Music BBS, Oceanside NY .. 516-536-8723, 24 hours, 1200/2400 baud PKWARE BBS ........................ 414-352-7176, 24 hours, 1200/2400 baud USENET ............................ katz@uwm-cs.milw.wisc.edu BIX ............................... Username: philkatz U.S. Mail ......................... 7032 N. Ardara Ave., Glendale, WI 53209 ------------------------------ Date: Monday, 21 September 1987 23:41-MDT From: kneller@CGL.UCSF.EDU (Don Kneller) Subject: NDMAKE 4.3 - make utility for MSDOS now available Now available from SIMTEL20... Filename Type Bytes CRC Directory PD:<MSDOS.C> NDMAKE43.ARC.1 BINARY 54214 30B6H NDMAKE (ver. 4.3) is a shareware make utility for MSDOS. For those un- familiar with `make', it is a utility that automates development of other programs. This MSDOS version includes: - auto response file generation to overcome the command line limit - VPATH (find sources in alternate directories) - full macro capability - default rules - looping over commands Don Kneller UUCP: ...ucbvax!ucsfcgl!kneller INTERNET: kneller@cgl.ucsf.edu BITNET: kneller@ucsfcgl.BITNET ------------------------------ Date: Sun, 27 Sep 87 15:50:10 EDT From: David Kirschbaum <kirsch@braggvax.arpa> Subject: HD Park software available from SIMTEL20 NetLandians, Recently read a query about PD disk park utilities on this newsletter. I'd suggest snarfing (all from SIMTEL20): PD:<MSDOS.DSKUTL> PDTIMPRK.ARC 9 kb TIMEPARK.ARC 4 kb TIMPARK1.LBR 4 kb and maybe?? DISKP.EXE 8 kb I snarfed one of these a while ago, works just fine, but can't remember which one. Haven't looked at the contents lately to see what's there, but expect nice assembler source (right?). Coulda sworn there was one clean utility to park on command, but can't find it. David Kirschbaum Toad Hall kirsch@braggvax.ARPA ------------------------------ Date: 28 Sep 87 11:35:23 GMT From: hsi!tankus@uunet.UU.NET (Ed Tankus) Subject: Commercial Programs Redirecting LPT# to disk There is a commercial product called PrintQ that retails for $89. I think it does exactly what you are looking for. It allows any print program to be spooled to disk for printout at a later date. The product has received favorable reviews from PC MAG and others. The company is in NJ. I don't have a phone number but they advertise in most PC magazines anyway. Cheers! -- Ed. Net : {uunet,ihnp4,noao,hao,yale}!hsi!tankus Snail: Health Systems Int'l, 100 Broadway, New Haven, CT 06511 Bell : (203) 562-2101 [The program LPTX.ASM in the Info-IBMPC Lending Library also works, is free, and is source code. --gph] ------------------------------ Date: Fri, 25 Sep 87 15:55 CET From: Helmut Waelder <ZRWA001%DTUZDV1%DEARN.BITNET@wiscvm.wisc.edu> Subject: PRTEGA.ASM I have added code to reset the printer to 6 lpi after graphics output. The complete source file follows: [PRTEGA.asm has been updated in the lending library - gph] ------------------------------ Date: Wednesday, 23 September 1987 07:17-MDT From: mcvax!dutrun!dutesta!karel@SEISMO.CSS.GOV (Karel van Houten) Subject: Unix compatible tools source and binarys from SIMTEL20 Now available from SIMTEL20... Filename Type Bytes CRC Directory PD:<MSDOS.DIRUTL> TOOLS.ARC.1 BINARY 69883 3B14H Some time ago there was a request for a UNIX compatible 'mv' command for MS-DOS. As a UNIX addict, I have written some commonly used UNIX tools in Microsoft C / MASM for myself. Here are the executables and sources for those who want to try them too. This package consist of: df, du, ls, mv, rm. -- Karel van Houten BITNET : NO! NO! NO! Delft University of Technology INTERNET : karel@dutesta.UUCP Faculty of Electrical Engineering UUCP : ..!mcvax!dutrun!dutesta!karel Room 9.29H, Mekelweg 4 SURFnet : DUTRUN::KAREL 2628 CD DELFT, The Netherlands VOICE : +31 15 783502 ------------------------------ Date: Mon, 28 Sep 87 21:23 EDT From: Hess.Unicorn@MIT-Multics.ARPA Subject: CapsCtrl & Buf160 Utilities, for new keyboards Mailed under separate cover for the Info-IBMPC library are two pieces of source code, both intended for effective use of the new almost-not- braindamaged IBM keyboards. The first, CAPSCTRL.ASM, makes the Caps Lock key turn into a control key. To get a "real" Caps Lock press, do Shift-CapsLock or Alt-CapsLock. [CAPSCTRL.ASM got lost in the Mail System Somewhere. gph] The second, BUF160.ASM, is a replacement for the existing one. It works with the new keyboards, and enables the use of the F11 and F12 keys as well as the usual 160 character buffer. No ill effects from using this with an old keyboard. Both CapsCtrl and the Buf160 changes are courtesy of Bill Spitzak at Mark of the Unicorn, Inc. Brian [BUF160NEW.ASM has been added to the Info-IBMPC Lending Library - gph] ------------------------------ Date: Sat, 26 Sep 87 09:26:57 EDT From: David Kirschbaum <kirsch@braggvax.arpa> Subject: Crossword Puzzles NetLandians, A critic recently asked me, "Well, if your computer is so smart, how come it can't do crossword puzzles?" Got into a discussion, suddenly realized I'd NEVER seen any code around to have a micro MAKE a crossword puzzle! Shouldn't be TOO hard .. especially in one of the recursive-friendly lan- guages like LISP or something, I'd think .. can anyone point me to some public domain source (any language, I guess) that creates crossword puzzles from word lists or dictionaries? Thanks in advance, David Kirschbaum Toad Hall kirsch@braggvax.ARPA [A cross word puzzle generator (written in FORTRAN) appeared in an issue of "Dr. Dobb's Journal" in the late 70's. I have the program somewhere in one of the boxes from my numerous moves. If I can find it, I'll send it to you and the Lending Library. I also saw a program (also written in FORTRAN) in a somewhat later issue of Dr. Dobb's that generated hidden word puzzles but for the life of me, I don't know where it went to... Does anyone have back issues of Dr. Dobb's? gph] ------------------------------ Date: Sat, 26 Sep 87 23:27:50 EDT From: weikert@nadc.arpa (J. Weikert) Subject: Eagle-PC info needed I know someone who has acquired an old Eagle-PC computer and would like to put a standard 20 meg. hard disk system in it. He has not been successful so far. Does anyone know anything about these machines? ------------------------------ Date: Sun, 27 Sep 87 10:29 N From: <IUS%DACTH51.BITNET@CUVMA.COLUMBIA.EDU> Subject: Reformat 1.50 trashing logical disks under DOS 3.3 I have been using your excellent Reformat program for about one year or so without any major problems. (I do backup, but so far, this has not proven necessary) One thing though. I have found that REFORMAT runs quite well under DOS 3.3 without any glitchs. However, if I want to use that new feature of 3.3 FDISK to have two logical disks on one physical (let's say a 33MB+) disk and reformat either of those logical disks then it trashes the other (logical) disk. Any ideas? Fixes? regards, el Eberhard W. Lisse Department of Plastic and Reconstructive Surgery and Burn Unit Technical University Aachen, West Germany ------------------------------ Date: Mon, 28 Sep 87 21:51:11 CDT From: "Rich Winkel UMC Math Department" <MATHPG1%UMCVMB.BITNET@wiscvm.wisc.edu> Subject: FORTRAN 'interval arithmetic' compiler query Does anyone know of such a beast for the PC? For the record, such a com- piler automatically keeps track of computational error bounds as the errors propagate through arithmetic calculations, so that results can be derived in the form 'answer= 1.22323 +- .00002'. Thanks for any help! Rich Winkel (MATHPG1@UMCVMB.BITNET) ------------------------------ Date: Thu, 24 Sep 87 22:55:40 MDT From: dd@lanl.gov (Dan Davison) Subject: Program always reads from drive A: I have an application program, available only as a .EXE file, which always tries to read a data file from the A: drive. Is there a way to change this using DEBUG? I'd like to change this to the C: drive. This program is not a commerical program, and source code is not available so that I can fix this problem. Any hints would be welcome. Dan Davison Arpa: dd@lanl.gov BITNET: dd%lanl.gov@wiscvm.bitnet CIS: 74065,41 UUCP: dd@lanl.uucp or ...cmcl2!lanl!dd Voice:505-665-1355| T-10 MS K710, Los Alamos Natl. Lab, Los Alamos,NM 87545 ------------------------------ Date: Thu, 24 Sep 87 22:55:58 MDT From: dd@lanl.gov (Dan Davison) Subject: How to add a 1.2 MB 5.25" floppy to an IBM PC/XT I have asked a local computer store if it is possible to add a 1.2 MB floppy drive to an IBM PC/XT (actually, it's an IBM PC Portable, which has the XT motherboard). After about a month they tell me this is not pos- sible. I'm aware that I'll have to get a different controller card, but surely there is a 1.2MB floppy controller for the PC bus. Does anyone have a high density floppy in an XT, and if so can you recommend a controller? Dan Davison Arpa: dd@lanl.gov BITNET: dd%lanl.gov@wiscvm.bitnet CIS: 74065,41 UUCP: dd@lanl.uucp or ...cmcl2!lanl!dd Voice:505-665-1355| T-10 MS K710, Los Alamos Natl. Lab, Los Alamos,NM 87545 ------------------------------ Date: Thu, 24 Sep 87 16:57:57 EDT From: Glen Jones <gdj5t@krebs.acc.virginia.edu> Subject: Adding a 3.5 inch drive to AT&T 6300 Is there any way to add an external 3.5 inch drive to an AT&T 6300 PC ***AND*** keep both internal 5.25 floppy drives? If not, are there any cautions involving buying/installing an internal 3.5 in place of a 5.25 inch? I would really like to keep both 5.25 inch drives, however. I am currently running PC-DOS 3.2 (by IBM) on the 6300; has anyone done this with a 3.5 inch drive on an AT&T6300? Details on my current setup follow: ROB bios 1.43 640K RAM HardCard20 installed (from Plus Dev.) both serial and parallel ports used Thanks in advance for ANY advice! Glen Jones U.Va. Biochem Dept., Box 440 > UUCP: ...!uunet!virginia!gdj5t Charlottesville, VA 22908 > BITNET: gdj5t@virginia.BITNET (804) 924-2373 > CSNET: gdj5t@acc.virginia.edu ------------------------------ Date: Fri, 25 Sep 87 17:03 EST From: <JWS%PSUARLC.BITNET@wiscvm.wisc.edu> Subject: Sunshine EPROM Burner We purchased a Sunshine EPROM burner through third parties. The accompany- ing software does not function on an EGA, but works OK on CGA. Do you know an address or telephone number for this company or a "quick fix" for the problem? Thanks in advance, Jack Sharer, JWS@PSUARLC ------------------------------ Date: Tue, 29 Sep 87 23:11:42 EDT From: Dimitri Vulis <CLDLV%NASAGISS.BITNET@WISCVM.WISC.EDU> Subject: Measuring speed from Turbo Pascal-bug fixed > From: RECK%DBNUAMA1.BITNET@wiscvm.wisc.edu (Gisbert W.Selke) > Subject: Measuring speed from within Turbo Pascal > > I need to measure the approximate clock rate of a PC from within a Turbo > Pascal program. So I wrote a little procedure which first got the cur- rent > time (via a DOS call), then executed some silly little loop and then got > the time again. > > >>> First question: Is there a simpler or more intelligent way to find the > speed? Consider the following: program measurespeed; function speed:byte; const NN=3; begin inline( $1E (* push ds *) /$31/$C0 (* xor ax,ax *) /$8E/$D8 (* mov ds,ax *) /$8A/$1E/>$46C (* mov bl,46c *) /$88/$DF (* mov bh,bl *) /$80/$C7/<NN (* add bh,NN *) /$3A/$1E/>$46C (*L1:cmp bl,46c *) /$74/$FA (*jz L1 *) /$40 (*L2:inc ax *) /$3A/$3E/>$46C (*cmp bh,46c *) /$75/$F9 (*jnz L2 *) /$1F (*pop ds *) /$88/$66/$04 (*mov bp+4,ah *) ) end; begin writeln('Speed returns ',speed); end. ALGORITHM: program measurespeed; const NN=3; var z:byte; count:integer; speed:byte; begin z:=mem0:$46C; (* or use 'absolute'... *) while z=mem0:$46C do; (* here we start a new tick *) z:=(mem0:$46C+NN) mod 256; (* wait for this value *) count:=0; repeat count:=count+1; until z=mem0:$46C; speed:=count div 256; end. EXPLANATION: On every PC-compatible, the timer interrupt (IRQ0) occurs 18.2 times per second and the handler routine increments the 4-byte value at 0:46Ch. We get its least significant byte, wait until it changes, and then increment a counter until the timer reaches t+NN, where NN is a number like 3 or 5. We return the MSB of the counter (i.e. counter mod 256) only, since the LSB is sometimes inconsistent (if a keyboard interrupt etc occurs, it steals some CPU time I guess). ADVANTAGES OVER PASCAL CODE: 1) The counter has to be 16-bit. When TP allocates it, it's as likely as not to fall on odd address, with significant performance difference on a 16-bit machine, not every time you run it, but when you change the program someplace else, so the address changes between even and odd. (the only real advantage) 2) The instruction mix generated for the repeat-until is a lot richer than loop L2, making the number of cycles more CPU model dependent. 3) On a very fast machine, the counter may overflow. We treat AX as un- signed. RELIABILITY: This function seems to give consistent results every time. This result is proportional (with a large error margin) to some sort of performance coef- ficient. Obviously, the result is affected by prefetch queues, wait states etc a whole lot. It's not really the clock speed, but some number propr- tional to some sort of 'performance index'. It's about as (un)reliable as Norton SI. ------------------------------ Date: Sun, 27 Sep 87 17:12:26 PDT From: Ya'akov_Miles%UBC.MAILNET@MIT-Multics.ARPA Subject: 80386 motherboard upgrade to IBM-PC/xt On Page 36, of "Computing Now!", Sept. 1987, is advertised an 80386 mother- board that is a drop-in replacement for the standard IBM-PC/xt motherboard. This 80386 motherboard runs at 6/16 mHz with zero wait states, and features four 16 bit and four 8 bit slots. Made by: Chie Dar Electronic Ind. Co., Ltd. Telex 34181 P.O. Box 89-36, Taipei, Taiwan, ROC Phone (02) 981-4487 Does anybody have any further information - it may be a cheap and dirty way to upgrade a plain vanilla XT-clone. A super "BABY-AT"? ------------------------------ Date: Tue, 29 Sep 87 13:06:45 edt From: moss!cbosgd!neoucom!wtm@RUTGERS.EDU (Bill Mayhew) Subject: Adding Memory to an IBM EGA Card One of our faculty people would like to increase the memory on his real IBM brand EGA card from 128K to 256K. He has the daughter board for holding the additonal chips. Alas, the board was installed by a dealer who was not kind enough to send along the installation notes. It looks as if the chips are just plain old 4164, 150 nS chips, but I want to be sure before I take screwdriver in hand (the IBM-supplied chips have some unrecognizable in-house part no.). When I add the extra memory, are there any jumpers that need to be changed? Thanks much to anybody who can assist. --Bill Bill Mayhew, Electrical Engineering Division of Basic Medical Sciences Northeastern Ohio Universities' College of Medicine Rootstown (what a name!), OH 44272-9989 USA phone: 216-325-2511 (wtm@neoucom.UUCP ...!cbosgd!neoucom!wtm) ------------------------------ Date: Tue, 29 Sep 87 16:51:00 MEZ From: Z3000JD%AWITUW01.BITNET@wiscvm.wisc.edu (Johannes Demel Digitalrechenzentrum) Subject: Experience with RightWriter/Rightwords Does anyone has experience with the Grammer and Style-Checker RightWriter and Dictionary-Utility RightWords from RIGHTSOFT Inc. (At the moment of- fered by IEEE at a reduced price). I am most interested using this in an application with the TeX and LaTeX Typesetting system by Donald Knuth. Johannes Demel, Technical University Vienna (BITNET: Z3000JD@AWITUW01) ------------------------------ Date: Tue 29 Sep 87 14:53:04-PDT From: Ted Shapin <BEC.SHAPIN@ECLA.USC.EDU> Subject: Uncompress for unix .Z wanted Phone: (714)961-3393; Mail:Beckman Instruments, Inc. Mail-addr: 2500 Harbor Blvd., X-11, Fullerton CA 92634 I would like to uncompress some unix .Z files. Is there such a program that runs under PC-DOS? ------------------------------ Date: Wed, 30 Sep 87 14:25:44 MEZ From: "Rainer Kleinrensing" UNI217%DBNRHRZ1.BITNET@WISCVM.WISC.EDU Subject: MultiSync monitor with Hercules Some time ago I bought a QuadEGA ProSync graphics card and a Multisync monitor from Quadram. I have read several times that Multisync compatible monitors can operate with a Hercules monochrome graphics card. The ProSync card is capable of emulating a Hercules card. My question: how do I have to connect the monitor to the card in order to operate in Hercules mode? The manuals don't give any clues. I have heard that all I have to change is the monitor cable. If anybody out there successfully operates a Her- cules card with a MultiSync monitor, please let me know. Thank you , Rainer ------------------------------ End of Info-IBMPC Digest ******************figu!re