Info-IBMPC@WSMR-SIMTEL20.ARMY.MIL (02/22/89)
Info-IBMPC Digest Tue, 21 Feb 89 Volume 89 : Issue 27 Today's Editor: Gregory Hicks - Chinhae Korea <COMFLEACT@Taegu-EMH1.army.mil> Today's Topics: 80387 in windows 386 A Question about CGA, EGA Enhanced keyboards Re: Questions on BBS' Turbo Pascal 3.0 -> 4.0 absolute cseg Harddisk on Olivetti M24 / AT&T PC 6300 HyperText Availability Internal format of .FWx files Memory parity interupt error on a Zenith AT (IBM clone) millisecond timer for PC Multi-tasking Software for IBM-XT Multitasking and modems under DesqView and Windows New msdos files uploaded to Simtel20 Non-standard formatting on IBM Diskettes Random numbers are more than flat distributions scanners and OCR software Memory Parity Interrupts ---------------------------------------------------------------------- Date: 17 Feb 89 13:37:00 CST From: zielke@physics.rice.edu Subject: 80387 in windows 386 I have recently been testing MS windows 386. There seems to be a quirk in the system that it does not save the 80387 registers during task switch. Is this a general problem with the system or have I forgotten to set this up. I know of several people running into this problem. Thanks in advance. David Zielke =============================================================================== ARPA==> Zielke@Physics.Rice.Edu * Zielke@128.42.9.23 * After Three Days MaBell==> 713-527-8101 ext. 4018 work * Without Programming 713-666-2982 home * Life Becomes US Snail==> David M. Zielke * Meaningless 7490 Brompton #110 * Houston, Tx 77025 * -The Tao of Programming =============================================================================== ------------------------------ Date: Sat, 18 Feb 89 21:08:46 EST From: "Robert E. Zaret" <ZARET@MITVMA.MIT.EDU> Subject: A Question about CGA, EGA Steve Gibson gave a particularly clear discussion of the various graphics standards in his InfoWorld columns in the 12 September to 24 October 1989 issues. ------------------------------ Date: Wednesday, 18 January 1989 13:31-MST From: Rich Kennerly <LVL@CORNELLA.CIT.CORNELL.EDU> Subject: Enhanced keyboards Here is a list of the extra keys on an enhanced keyboard: Enhanced Keyboard key: BIOS value: base shift cntrl alt F11 8500 8700 8900 8b00 F12 8600 8800 8A00 8C00 kp-Enter E00D E00D E00D A600 KP-/ E02F E02F 9500 A400 KP-* 372A 372A 9600 3700 HOME 47E0 47E0 77E0 9700 UP ARROW 48E0 48E0 8DE0 9800 PAGE UP 49E0 49E0 84E0 9900 LEFT ARROW 4BE0 4BE0 73E0 9B00 RIGHT ARROW 4DE0 4DE0 74E0 9D00 END 4FE0 4FE0 75E0 9F00 DOWN ARROW 50E0 50E0 91E0 A000 PAGE DOWN 51E0 51E0 76E0 A100 INSERT 52E0 52E0 92E0 A200 DELETE 53E0 53E0 93E0 A300 Note that you cannot get these values with the old BIOS interrupt, it will remove the E0's to be compatible with the old keyboard. See the following code fragments that use the old and new BIOS calls. getkey_ PROC NEAR ; get key code for standard keyboard mov ah,01H int 16h jnz GOTAKEY mov ax,00H ret GOTAKEY: mov ah,00H int 16h ret getkey_ ENDP getxkey_ PROC NEAR ; get key code for enhanced keyboard mov ah,11H int 16h jnz GOTXKEY mov ax,0 ret GOTXKEY: mov ah,10H int 16h ret getxkey_ ENDP I have found that an AT with old keyboard and old BIOS (you need the new BIOS to fully appreciate the enhanced keyboard) will get very upset if you try to use the new BIOS function calls with it. Here is a routine that will look into the BIOS data area to see if BIOS found an enhanced keyboard at INIT time. Note that this approach is not ideal since IBM and others can change the layout of BIOS at will; fortunately they never seem to. In any case there is no interrupt call that will tell you this (if there is let me know please). ; gkybdtype returns 0 for standard 84 key kybd, 10H for enhanced keyboard KB_FLAG_3 equ 096H ; address of keyboard flag byte KBX equ 00010000B ; enhanced keyboard flag bit gkybdtype_ proc near push es mov ax,40H mov es, ax mov ax,es:[KB_FLAG_3] and ax,KBX pop es ret gkybdtype_ endp RICH KENNERLY - LVL@CORNELLA.CIT.CORNELL.EDU CIT - Network Development 125 Caldwell Hall 607-255-7342 ------------------------------ Date: 17 Feb 89 20:49 +0100 From: Dag Baardsen <daghb%sun.dhhalden.uninett@norunix> Subject: Questions on BBS' I am looking for some BBS-software (Public Domain) that runs on an Ethernet (3Com) network. Are there anybody who could help me with this topic? In case you do, is the source code included? Yours... Dag H. Baardsen uninett, norway ------------------------------ Date: 17 FEB 89 21:22:15 CST From: John Stewart <JSTEWART@SFAUSTIN> Subject: Turbo Pascal 3.0 -> 4.0 absolute cseg References: John Stewart <jstewart@sfaustin.bitnet> I have recently come across several programs written in Turbo Pascal 3.1 that I am interested in using. I have Turbo Pascal 4.0, and have found a conversion problem that has gotten me puzzled. In the 4.0 manual is states: In version 3.0, you could use CSeg and DSEG in absolute statements: var Parameters: string[127] absolute CSeg: $80; In version 4.0, neither CSeg nor DSeg is allowed in absolute statements. That is all they have to say about it. Can someone show me how I can get around this problem? Why did they remove this capability? I assume there is a logical reason, but I have never used these two statements before and am not familiar enough with the commands to find another way around it. If anyone could possibly give me an example of how the above code could accomplish the same task, and still compile under 4.0 I would be greatly appreciative. Send replies to <jstewart@sfaustin.bitnet> . I will send a reply back if/when I figure it out. Thanks, John Stewart Academic Support Programmer Stephen F. Austin State University Nacogdoches, Tx 75962 (409) 568-1020 voice (409) 568-1334 BBS (data) ------------------------------ Date: Fri, 17 Feb 89 15:12 UT+1 From: "GERBEN POSTMA" <DDGMPOS%HDETUD51.BITNET@CUNYVM.CUNY.EDU> Subject: Harddisk on Olivetti M24 / AT&T PC 6300 Does anyone know if it's possible to connect a HD controller or hardcard directly to the motherboard bus (to which one otherwise should connect a busconvertor)? I've got an old Olivetti M24 / AT&T PC 6300 that is not equipped with a busconvertor. Now I wonder if I can connect, using e.g. some flat cable and appropriate connectors, a controller or hardcard directly to the motherboard bus, rather than buying a quite expensive busconvertor giving me 7 slots, of which I need only one... Would my BIOS Rev 1.1 be OK? Aloha & mahalo! Gerben Postma Delft University of Technology Bitnet: DDGMPOS @ HDETUD51 ------------------------------ Date: Sat, 18 Feb 89 21:16:22 EST From: "Robert E. Zaret" <ZARET@MITVMA.MIT.EDU> Subject: HyperText Availability The 30 January issue of InfoWorld (page 8) has an article with the title "IBM Ships Hypercard-Like Product for Multimedia". The product is called Linkway. It supports color. The minimum environment has one 360K floppy, 384K of memory, and a mouse. It is not a direct competitor for Hypercard. Another product, called LINKER, is available from Pugh-Killeen Associates 1 Bowdoin Street Newton Highlands, MA 02161 I don't know much about it, although I have a demo that fits on a single 360K floppy. ------------------------------ Date: Tue, 14 Feb 89 17:51:26 +0100 From: mcvax!crg.bull.fr!Stavros.Macrakis@UUNET.UU.NET Subject: Internal format of .FWx files Does anyone have a document or a program documenting the format of Framework's FWx files? I would like to recuperate text, formulae, and data, as well as documents' tree structure. Thanks. -s Net: Stavros.Macrakis@crg.bull.fr or ...!mcvax!inria!bull!crg!Stavros.Macrakis ------------------------------ Date: Wed, 15 Feb 89 17:43:24 +0100 From: Andre' PIRARD <A-PIRARD%BLIULG11.BITNET@CUNYVM.CUNY.EDU> Subject: Memory parity interupt error on a Zenith AT (IBM clone) The IBM AT advanced diagnostics are very clever at detecting an out-of-specs memory chip and tell you which when all other tests (even the plain diagnostics) will just make you feel happy. Assuming good memory and DMA chips, I'd check the device. While the DMA is responsible to present the addresses, the device is to output the data and may fail to do that correctly (unstable signals). Andr. ------------------------------ Date: Thu, 16 Feb 89 14:21:13 +0100 From: Johan Van Rensbergen <FPAAS01%BLEKUL11.BITNET@CUNYVM.CUNY.EDU> Subject: millisecond timer for PC In Info-IBMPC Digest Vol 89:Issue 19 G. Nassas asks about millisecond timing on an IBM-PC. I hope this small C-program helps. More on this matter can be found in Behavioral Research Methods, Instruments & Computers, a magazine from the Psychonomic Society, which probably is available in all psychology-libraries. Johan Van Rensbergen Lab. Experimental Psychology University of Leuven, Belgium. e-mail : FPAAS01 at BLEKUL11.bitnet Phone : +32-16-285960 /* * timerprogram * * J. Van Rensbergen * K.U.L., University of Leuven, Belgium * Dept. of Psychology */ #include <dos.h> #include <conio.h> #define intvect 0x1c #define port_c3 0x43 #define port_inp 0x40 int counter, i, kk; void interrupt newint() { ++counter; } void fast() { outportb(port_c3, 0x36); outportb(port_inp, 0xa9); outportb(port_inp, 0x04); } void slow() { outportb(port_c3, 0x36); outportb(port_inp,0x00); outportb(port_inp,0x00); } main() { void interrupt (* oldint)(); clrscr(); oldint = getvect(intvect); setvect(intvect, newint); printf("begin\n"); fast(); counter = 0; while(!kbhit()) {}; i = counter; slow(); kk=getch(); setvect(intvect, oldint); printf("counter = %i",i); } ------------------------------ Date: Wed, 15 Feb 89 18:55:00 +0100 From: Andre' PIRARD <A-PIRARD%BLIULG11.BITNET@CUNYVM.CUNY.EDU> Subject: Multi-tasking Software for IBM-XT If your CPU is a 386 and you're not frightened by slightly technical details, you may trust DESQview+QEMM386, especially for an application oriented environment. Your (real) memory size will be the limit. One has to *physically* disconnect main memory and install EEMS with other processors. Once installed correctly, well behaved applications will run smoothly and the only user concern will be to not open the same file twice. And even SHARE is supposed to take care of that! If you wish to move most of TSRs, you may *in* applications, else it reduces virtual memory space. Nut with a 386, these guys from Quarterdeck even throw prerequisite code (even device drivers) up high above 640 Kb (where not all are happy to be, by the way... unhappily. (Oops!). Maybe a Windows *386* could be a challenger, but I guess it will *eat* your memory just as much as the other Windows versions (DV is happy in 140 Kb). I may finally sound fanatic, and I confess my AUTOEXEC.BAT ends with DV and my DV.BAT with PARK, but I am a extremely delighted DV user first of all. And I prefer not spending my time sparing money for OS/2, even though it looks like I've got plenty of time. Andr. ------------------------------ Date: Wed, 15 Feb 89 18:06:28 +0100 From: Andre' PIRARD <A-PIRARD%BLIULG11.BITNET@CUNYVM.CUNY.EDU> Subject: Multitasking and modems under DesqView and Windows At 4.77 Mhz, DESQview will start to fail above 9600 bauds in the foreground. At higher CPU speeds, it's generally OK. As I rarely use terminal mode in the background :-), I can only tell from observing file transfer retry counts. You're probably fine for several 2400 bauds lines on an AT (especially if in the same application). Your BBS should not write directly to the screen (if it's supposed to write anything *there*), else you'll need to either modify it to play the DESqview/Topview/Windows game or use a 386 and QEMM. The real danger does not come from DESQview itself. It's from how well your other applications (that includes DOS (generally safe) and your BIOS) behave at not disabling interrupts and looping for fun. I wouldn't bet a cent on Windows 1, nor on 2 despite I never really tried it. Andr. ------------------------------ Date: Sat, 18 Feb 1989 01:32 MST From: Keith Petersen <w8sdz@WSMR-SIMTEL20.ARMY.MIL> Subject: New msdos files uploaded to Simtel20 [--forwarded message--] From: madler@Hamlet.Caltech.Edu (Mark Adler) I have uploaded 10 .ARC file to Simtel20. The files are: pd1:<msdos.turbo-c> TCRND.ARC High speed, high quality random number generators for Turbo C. pd1:<msdos.filutl> ASCIFY13.ARC Converts binary files to text and back for transfer over bitnet or other unfaithful channels. pd1:<msdos.dirutl> NEWCHDIR.ARC A new change directory command (c.com) that accepts abbreviations. pd1:<msdos.dirutl> SZ14.ARC A command that determines the number of bytes and amount of space used by files and directories. Can also search for files, find unbacked up files, etc. pd1:<msdos.dirutl> CA12.ARC A change attribute command---works over entire subdirectories if needed. pd1:<msdos.dirutl> MA_RM13.ARC A powerful delete command that can remove entire subdirectories. It will delete without prompting for use in batch files, or can prompt on a file by file basis. pd1:<msdos.ega> EGACOLRS.ARC Displays all 64 colors at once on an EGA to aid in palette selection. <pd1:<msdos.ega> FIX43.ARC TSR to fix the cursor in EGA 43 line mode. pd1:<msdos.ega> ZAPSCRN.ARC TSR's to blank the screen after a period of inactivity for the EGA (egazap.com) or VGA (vgazap.com). pd1:<msdos.keyboard> KBUF11.ARC TSR to extend the size of the keyboard buffer. All of these .arc files contain documentation, source, and executable files (or .obj files in the case of tcrnd). Mark Adler bitnet: madler@hamlet arpa: madler@hamlet.caltech.edu Mark: thanks for a very nice collection of original programs. You certainly have been active! --Keith ------------------------------ Date: Fri, 17 Feb 1989 19:57:22 PST From: Billy Brackenridge <billy@venera.isi.edu> Subject: Non-standard formatting on IBM Diskettes See Volume #1 issue #1 of INFO-IBMPC back in August 1982 for a discussion of this subject. Back in the DOS 1.1 days when hard disks cost about $5K we had to make the best we could from floppy disks. Tall Tree systems made a package called JFORMAT that would format 360K disks (properly 440K disks) in ten sectors. They also supported Quad density disks in 880K format. Please don't call them they dropped this line of software when DOS 3.0 was announced and I think only John Henderson, the company president remembers this episode in the history of MS-DOS. Yes, it can be done but it shouldn't. IBM is overly conservative with 8 sectors, but from experience in this area I can tell you 10 sectors is pushing the technology (or at least was when a floppy disk cost $400). ------------------------------ Date: 17 Feb 89 10:24 From: dantowitz%eagle1.DEC@decwrl.dec.com (This space intentionally filled in) Subject: Random numbers are more than flat distributions Note that a "random" generator that generated the following sequence: 1, 2, 3, 4, 1, 2, 3, 4, ... would have a uniform distribution but would be far from random. A true random number generator must generate all pairs, triplets, quads, etc. with the same frequency. In other words the sequence 1, 2, 3 must be presented as often as 1, 3, 2, and all the other permutations. This is very important when simulating a series of events based on multiple calls to a random number generator. ------------------------------ Date: 17 Feb 89 10:19:44 EDT From: JBAUER@BAT.Bates.EDU Subject: scanners and OCR software I'm interested in anyone having good information about desktop scanners. We're looking for a flatbed unit with good gray scale capabilities as well as saftware for OCR. Especially typeset text. ( the problem ) We're using Word Perfect (5.0) as well as Ventura and Publishers Paintbrush. Any info would be greatly appreciated. Jim JBauer@Bat.Bates.EDU ------------------------------ Date: Fri, 17 Feb 89 09:54:35 PST From: swillett@plutonium.cchem.Berkeley.EDU (Steve Willett) Subject: Memory Parity Interrupts This is in response to a query which appeared in a recent issue of Info-IBMPC Digest. I have mailed the response to the person who posted the query so if it is not of general interest please redirect to /dev/nul. I saw your note in Info-IBMPC Digest concerning memory parity interrupts. We have had considerable problems with the combination of Data Translation A/D/A boards (DT2801) and Ungermann-Bass network boards in the "true blue" IBM AT's we are using in an undergraduate process control lab. There is apparently some kind of intermittent, subtle timing problem between these two boards which has caused various symptoms on different machines at different times with some boards (vague enough for you?). Most of these problems have been eliminated by replacing the U-B board with an apparently identical board. The replaced board is then used elsewhere in the Colege of Chemistry with no problem - so it seems to be the combination of Data Translation and Ungermann-Bass. This may be of interest to you. Are you using any other boards in your clone which use DMA? A network board? One of the errors we got on another machine was undetected errors in the digital signal stored from the A/D/A board in main memory. This showed up as "sticky bits", i.e., the digital data had some bit pattern which would not change as the data changed. If we accessed the A/D/A board directly (no DMA) the data was fine. If we changed the A/D/A board the error persisted. The only way we were able to make this one go away was to remove the network board entirely - whether or not we started up the network connection had no effect. I don't know if this is any help, but good luck! Steve Willett swillett@plutonium.cchem.berkeley.edu ------------------------------ End of Info-IBMPC Digest ************************ -------