Info-IBMPC@B.ISI.EDU.UUCP (07/17/86)
Info-IBMPC Digest Wednesday, July 16, 1986 Volume 5 : Issue 70 This Week's Editor: Richard Gillmann Today's Topics: TMS320 Signal Processing Cards AT->PC Disk Compatibility (2 msgs) TeX (2 msgs) WordStar EGA Patches TEKTERM, PUSHDIR/POPDIR PUSHDIR/POPDIR Added to Library GDIR, REMIND, SETTIME, ERROR Added to Library SCRNCOLR.ASM Added to Library LEX && YACC Interactive EasyFlow Simula BASIC Compiler File SMALLERR.OBJ PC Net 1.0 vs PC-DOS 3.2 Today's Queries: Hard Disk, Plotter Queries ATACK AT Motherboard Query Problems with New IBM PC/XT Display-Write 3 File Recovery EMS Board Query Special Multi-Function Card Wanted ---------------------------------------------------------------------- Date: 15 Jul 1986 19:11:20 PDT Subject: TMS320 Signal Processing Cards From: Billy <BRACKENRIDGE@B.ISI.EDU> To: info-ibmpc@B.ISI.EDU For the last few years I have been working with several TMS320 based signal processing cards for the IBM PC. The TMS320 is a family of high speed signal processor micro processors capable of a few million multiply/add operations per second. There are several cards on the market that combine the TMS320 with analog to digital converters and telephone interfaces allowing software implementation of modems, telephone control, and speech recognition/compression. In addition TMS320s can be used for robotics, image processing, data encryption, fax encoding, or any bit manipulation or arithmetic intensive operations. TMS320 based boards for the PC are made by ITT, Natural Microsystems (Watson), Texas Instruments, and IBM as well as numerous small companies making versions appropriate for laboratory use. I have had first hand experience with the Texas Instruments and IBM cards and thought it worth passing along some observations. If you are interested in buying a card as a consumer and just using it in your PC as a modem, answering machine, and telephone dialer, the Watson card from Natural Microsystems is probably your best bet. It sells for about 1/3 the cost of the other cards and the user software is reputed to be pretty good. I haven't had any first hand experience with Watson as there was no programmer's interface when I last enquired. Currently Watson does modem, telephone dialing, and speech play/record. Text to speech and speech recognition could be added as these can be run mostly in the PC. As the Texas Instruments and IBM systems cost around $1200-$1500 I can't see them replacing your $39.95 K-Mart answering machine, but the Texas Instruments board wins for a clean easy to use out of the box interface. The TI answering machine and telephone dialer is combined in one resident program that pops up with a touch of a key. The IBM answering machine application code expects to run under Topview. When run in the normal DOS environment it takes over the entire screen and processor. The phone dialer is a separately priced program. It stays memory resident like the TI program, but can't interact with the IBM answering machine. I have used both the TI and IBM systems as my answering machine. Both offer more features than a sane man would want and neither have the user interface I would choose. Both maintain elaborate undocumented rigid file structures that are impossible to manipulate with any other program. Unless one is already running Topview or other window package I prefer the TI system as the lesser of two evils. I know of no modem software available for the TI system. I would like for TI or some third party to prove me wrong. IBM provides modem firmware for their card, but as it does not use the standard serial port interface, no communications programs are compatible. I wrote a simple glasstty program that uses the modem firmware to implement a dumb tty. This application also dials the remote computer and senses rings and busy signals etc. I have used it to connect to local bulletin boards. I have placed GLASSMODEM.ASM in the info-ibmpc lending library. I may write a similar interface to Kermit and give it to Columbia, but don't hold your breath. The rest of this article will be from the point of view of an applications programmer. Richard Gillmann, Andy Witkowski, and I recently completed a project called VTMS. VTMS stands for Voice Text Mail System. It consists of an AT equipped with the Texas Instruments speech card and an Ethernet interface. It has been sitting in a corner of the machine room at ISI for the last few months. Subscribers call VTMS from a touch tone phone. VTMS fetches their text mail from any host on the internet and uses TI's text to speech algorithm to read them their mail. Subscribers may also send and receive voice mail. This is just digitally recorded speech which goes nowhere but can be played out over the phone. Subscribers may also respond to text mail from non subscribers by sending them voice mail. The non subscriber will receive a text message from VTMS something like: Received: FROM VTMS-01.ISI.EDU BY USC-ISIB.ARPA WITH TCP ;10 Jul 86 23:26:33 Date: 10 Jul 86 22:28:25 From: brackenridge@B.ISI.EDU To: Info-IBMPC@B.ISI.EDU Subject: Re: Demo Message In-Reply-To: (Message from "Info-IBMPC@B.ISI.EDU" of 10 Jul 1986 23:16:01 PDT) In response to your text message of 10th of Jul. 86 Billy Brackenridge has left a voice message for you. Please dial (213)555-5555. When prompted for an access code use a Touch Tone phone to enter the access code 954. VTMS operator. This way someone can respond to text mail without using a terminal. IBM has a similar system they use on VNET called TAOS which is based on their speech card. The first such system of this nature was built by the MIT Architecture Machine Group and since then they have done many innovative speech applications. This is a whole new area that is just becoming practical, and one where people with imagination might be able to make their fame and fortune or at least have some fun. I hope others will write some speech applications and perhaps even donate something to our library! To this end a little technical discussion of the differences between the TI and IBM speech cards is in order as these are the two candidate cards one might use. Both cards feature the TMS32010. As these microprocessors are Harvard architectures, memory is divided into program and data RAM. IBM has a bit more program RAM and TI has more data RAM. I think IBM made the better choice here as their larger program RAM allows them to run an operating system on the TMS320 which allows context switching between up to four separate tasks. If you are interested in speech recognition, TI's choice of a larger data RAM allows room for more vocabulary templates. It's all engineering trade offs. The TMS32010 requires use of fast static RAMS, AS and ALS logic which are very expensive. The cost of the board rises as the size and sophistication of the memory increases. The Watson board is cheaper as it has far less memory available to the microprocessor. The next generation of boards use the TMS32020 microprocessor which can run with slower dynamic RAMS. This could drive the cost down and functionality up, but there doesn't appear to be enough demand for these cards to spur management to produce the cards currently running in the laboratory. All of these boards feature telephone interfaces and A/D converters. IBM is the big winner here! IBM has two pairs of A/D D/A converters that do DMA access to circular buffers directly in memory without microprocessor intervention. Although none of their software supports it, the IBM card is capable of servicing two lines at the same time. The IBM card also features dual sampling rates. It can sample at 8000 or 9600 samples per second. This portion of the card is a very clever design. The TI card uses a CODEC for A/D conversion. This means the microprocessor must convert linear values to log and visa versa. This is a bother and complication, but the real problem is that 8 bits log (equivalent to 12 bits linear) is not enough dynamic range to handle a phone line. CODECs work fairly well in telephone central offices as the signal levels of the phone line can be regulated by technicians. At the subscriber end there is nothing you can do if your phone voltage is too high or low. While IBMs linear converters are only 12 bits wide, they have added a programmable amplifier that adds another four bits to the dynamic range of incoming signal. This added dynamic range is very important when recording incoming speech from long distance phone lines. Experience with VTMS has shown that the TI system degrades when recording voice over long distance phone lines. This isn't quite as important for the telephone answering machine where the owner of the machine is the only one who hears the messages, but in the VTMS application we have found that people get a bad impression of the system if they are using it for the first time and the voice quality is not up to par. Another big difference between the two boards is how the various companies do touch tone detection. IBM uses digital filtering on the TMS320 to do it in software, and TI uses a real lousy touch tone decoder chip. It works most of the time, and I wish I could characterize when it doesn't work, but it fails enough to make life miserable. In VTMS we were able to work around this hardware deficiency by timing out and assuming a default key for the user. By careful arrangement of the command set one can live with flaky touch tone detection, but there is no excuse for bad hardware design like that. I haven't had much experience with IBM's algorithm, but I have worked with my own and as it was entirely written in software it could always be adjusted to work in all situations. Aside from the hardware limitations of voice record and play, Texas Instruments is clearly superior to IBM and Watson in their speech compression algorithms. TI can record at 2400, 9600 and 32000 bits per second where IBM records at 14,400, 19,200 and 28,800 bits per second. TI has done a great job in achieving high quality at low bit rates, however, the IBM algorithms have a better AGC. You can shout or whisper into the telephone and your recorded voice comes out at the same level. While this may not be appropriate for all applications, it should solve the problems we have with long distance phone calls in VTMS. Neither group has been able to master the art of silence detection. See Merritt below for a reference on how to do this. Both TI and IBM use the text to speech algorithm from Speech Plus. IBM's version appears to be about a year's more refined than TI's. IBM also integrated text to speech into their operating system better than TI did. A major portion of text to speech code must run in the PC. IBM is able to dynamically load it and then free up the space. With TI, once you have loaded text to speech, you can never get the 120K bytes back without rebooting. Unfortunately the Speech Plus algorithm uses a resonator model of the vocal tract. This can eat up 75% of the TMS320's cycles and limit the rate at which the card can speak. In IBM's case this means you can't run touch tone detection and text to speech at the same time. While IBM has provided limited touch tone detection algorithm which only recognizes the * tone, all this is bad engineering. If they had used a lattice filter representation of the vocal tract model such as Berkeley Speech Technology uses, they would only need about 15% of the TMS320 cycles, and the IBM card would be able to run two channels of text to speech as well as full touch tone detection on both channels. I can't say much about the relative merits of the speech recognition algorithms on the two cards, as I have done little more than play with the demos. TI uses their own dynamic time warp LPC template matching algorithm which runs almost entirely on the TMS320. IBM uses Dragon Systems' Markov process time domain stuff which runs mostly on the PC. Both systems have their good points, and anyone planning a project using speech recognition should investigate both systems and make their choice on criteria appropriate to their specific application. IBM really shines when it comes to documenting what they produce. TI has been pretty good in sending me internal documents so I can program their beast, but IBM hasn't had to. The Options and Adapters manual even includes schematics for the board. Suppose you want to write an application using the existing play, record, text to speech, phone dialing and call progression, and touch tone detection features of the two boards. If you have a simple application you want to get running quickly, TI is your best bet as there are well documented high level language interfaces that will get you running quickly, but lord help you if you don't want your interface exactly the way TI wrote it. They have some pretty demented ideas about how they think people ought to program. Also forget it if you want to run in a multi tasking environment. For VTMS we had to throw away all TI's code and interface directly to the hardware and firmware. I spent a month or so rooting around with the Codesmith debugger in TI's code, and it is disgusting. TI ought to learn to hire programmers not engineers. At first glance IBM's code is a little bizarre in that one is faced with a lot of IBMspeak. The documentation reads like an MVS program logic manual, but the code is a work of art. Everything is reentrant and asynchronous. IBM has correctly recognized that speech applications are well suited to multi tasking windowed environments and that a resource such as the speech card is best shared between several applications. While it takes a little longer to get started with the IBM programmer's interface, the richness of their low level command set is well worth the extra effort. When coupled with the superiority of their hardware phone interface, IBM is the only sensible choice for telephone based applications. All of these cards have been a dismal failure in the market place. I understand from PC Week that IBM is dropping further development in this field. As usual PC Week had their opinions mixed with facts and both were confused. It is unfortunate as IBM was on the right track from a technical point of view, but marketing support for this product was nonexistent. Clearly speech and signal processing applications are going to be increasingly important in the next decades. It seems stupid to drop the whole effort because the first generation card was badly marketed. I will conclude with a few ideas for applications that could be developed on this generation of board: I'd like to see a simple memory resident phone dialer for DOS that would read phone numbers from the screen from any program such as VDTE, Dbase or Epsilon. Specialized programs which insist on keeping phone numbers in weird file structures seem pretty silly to me. The program should be able to parse phone numbers in a variety of formats adding dialing prefixes and area codes as necessary. Danny Cohen wrote such a program on a 68000 and connected to a serial port on his HP terminal several years ago. I assume he still uses it with his PC when running the VDTE terminal emulator. I haven't spoken with Danny in several months so I don't know if he still has a PC or has moved to the brave new world of SUNs and unix. Nobody has written a Windows application for the IBM card. One could combine the phone dialer I mentioned above with a icon interface. I might even run my copy of Windows if there was such an application. Chris Schmandt and his colleagues at the MIT Architecture Machine Group have done some fun applications experimenting in human computer interaction. Their "phone slave" combines speech play and record, speaker identification, voice recognition, and a graphics interface in a manner that is really exciting. Three years ago this system required a dedicated super-mini computer and lots of extra hardware, but today it could be done on a PC. I understand a version of the phone slave is marketed by a company called Active Voice in Seattle. FAX and image processing is an area rarely touched on. Here at ISI there are two systems using TMS320s for image compression over the INTERNET. An interesting application is combining the software modem capabilities of the 320 with FAX encoding/decoding. The TMS320 is really spiffy at bit testing and inserting as it has a barrel shifter which makes FAX go really fast. FAX modems (at least outside of Japan) are expensive as they are specific to FAX only. The current generation of boards would be cost effective for this, but the market is limited. Why did IBM neglect the wider modem market and Texas Instruments ignore it entirely? Currently IBM has implemented only the 300 and 1200 baud modems. I am surprised that they never got to the 2400 baud or 4800 baud half duplex modems used for talking to IBM mainframes. If it talked to IBM mainframes, IBM sales people might get behind selling the Voice Communications Option. If I ever find the time, I plan to adapt the IBM card to amateur radio. Currently there are between twenty to thirty thousand packet radio nodes in the United States. With a one watt radio, one can use digipeaters to forward packets over a range greater than is possible with direct transmission. Three years ago I wrote a sub thousand bit per second LPC vocoder that ran on a TMS320 board Paul Mockapetris and I built for the PC. The idea is to send speech digitally in packets, but as the bandwidth of the packet net is pretty low one must do a radical encoding of voice. The quality of speech isn't much, but it is better than the single side band amateurs are currently used to. Such a system could be used with Tymnet's "PC Pursuit" offer of unlimited after hour data calls for $25 a month to provide unlimited voice calls for the same price. The quality would be adequate for multi user interactive computer games where voice could be mixed with data. Maze wars with voice anyone?? The whole idea of voice games hasn't been exploited much. I think there are possibilities to start a whole new industry here. Dick Gillmann wrote "Richard's Telephone Matchbox" last summer. It is a voice/touchtone dating service. Subscribers can send voice mail to one another, post public messages, and match to other users to meet people with similar interests. The program is still running at (213)676-7764. Crude rip-offs of this idea are running on some of the 976 numbers. (213)97-NITES will cost you 95 cents plus toll charges to meet the girl or boy of your dreams. These cards are finding use in the handicapped market. There are better speech recognition devices on the market, but none of them do text to speech. It seems to me the flexibility of a generalized signal processor connected to a microphone/speaker offers a lot more potential than a rigid hardware approach to speech recognition for the handicapped. Suggestions for further reading: Volume 4 issue 121 and 124 of INFO-IBMPC: Text of IBM announcement of Voice Communications Option and my comments. Volume 3 number 5 IBM Personal Computer Seminar Proceedings: Voice Communications Option IBM Voice Communications Application Program Interface Reference IBM Part No. 6280743 0743 IBM Technical Reference Options & Adapters: Voice Communications Adapter; 55X8864 October 11, 1985 TI-Speech Applications Toolkit TI Part No. 2249821-0001 Providing Telephone Line Access to a Packet Voice Network Ian H Merritt University of Southern California Information Sciences Institute, February 1983 ISI/RR-83-107 Summary Proceedings of the UCLA-NSF workshop on Personal Communications Izhak Rubin Editor UCLA-ENG-85-29 Richard Gillmann (Richard's Telephone Matchbox) p.353 Chris Schmandt (geek@nedia-lab.mit.edu) can be contacted for further information about the phone slave and other projects done at the Architecture Machine Group. [GLASSMODEM.ASM has been placed in the lending library -wab] ------------------------------ Date: Mon, 14 Jul 86 11:58:56 SET To: INFO-IBMPC@USC-ISIB.ARPA From: ESC1111%DDAESA10.BITNET@WISCVM.ARPA Subject: AT->PC Disk Compatibility A while back I asked about AT->PC disk transfers. All the answers have convinced me that the 'naive' approach should have worked and that one or other of the disk drives must be misaligned. I know how to checkout PC (360K) drives (various firms sell calibration disks) but I've never seen the same thing for AT (High density) drives. Any Ideas ? snail : N. Head, European Space Operations Centre RobertBoschstr. 5 6100 Darmstadt Federal Rep. Germany bitnet/earn : ESC1111 at DDAESA10 BIX : nhead ------------------------------ Date: Tue, 15 Jul 86 08:54:48 EDT From: SAROFF%UMass.BITNET@WISCVM.ARPA (MATTHEW G. SAROFF) Subject: AT->PC Disk Compatibility To: INFO-IBMPC@USC-ISIB.ARPA I have talked to two people who are knowledgable about IBM PCs about the 1.2 meg 5.25" floppies available. They said: 1) "There is not a problem. People are using cheap disks, and with the data packed in so tight, you get REAL problems with cheap disks.", AND 2) "There are some real problems. For the 1.2 meg format, it is only 60% likely that another drive will be able to read the disks, and if you use the 360K format, this goes down to less than 30%." Needless to say, this child is downright confused. What are individuals' experiences with ther 1.2meg drives? Matthew Saroff ------------------------------ Date: 150786 11.00.00 To: <INFO-IBMPC-REQUEST@USC-ISIB> From: <ACSKATJ%HUTRUU0.BITNET@WISCVM.ARPA> Subject: Re: TeX In Notices of the American Mathematical Society (Volume 33, No. 1, Jan. 1986) an article about Technical Word Processors was published. In that article a number of TWP's are compared and also some TEX- implementations for PC's are described. Jan van Kats Academic Computer Center Utrecht State University of Utrecht Budapestlaan 6, "de Uithof" Utrecht, the Netherlands EARN: ACSKATJ@HUTRUU0 UUCP: MCVAX!ACCUMV!ACSKATJ ------------------------------ Date: Tue 15 Jul 86 11:28:31-PDT From: David R. Fuchs <FUCHS@SU-SCORE.ARPA> Subject: Re: TeX To: INFO-IBMPC@B.ISI.EDU To correct some mis-information in the last posting about TeX: (1) The MicroTeX base price is $295; with an Epson driver, $395. This has been true for the better part of a year now. (2) LaTeX runs fine under MicroTeX. The LaTeX manual is included in the LaTeX box that is optionally available. (3) The MicroTeX Epson driver runs faster than the PC-TeX one, when running at the same quality settings. MicroTeX's Epson driver also has more speed-vs.-quality options. (4) The MicroTeX LaserWriter driver runs just the same speed as the PC-TeX one (since it's actually the same program, licensed from Textset, Inc.). Even more importantly, MicroTeX 1.5 is significantly faster and smaller than PC-TeX. The speed has been measured at 25% to 33% faster, depending on the job. The small size, in conjunction with the internal "poor-man's virtual memory" scheme means that MicroTeX can handle much larger and complex TeX jobs. You get more room for macros, fonts, pages, etc. This is especially important when running macro packages like LaTeX. The only space-hogging aspect of either version is the font-image storage. If you are only using a few fonts, you can get by with well under 1Meg of disk space (indeed, MicroTeX even runs on a dual-floppy system). Finally, the Tex Users Group yearly meeting is July 21-23 (at Tufts Univ.), and A-W will be demonstrating some major new functionality in MicroTeX that I will be glad to report on the following week. -david (who decidedly has an interest in MicroTeX) ------------------------------ Date: Tue 15 Jul 86 11:56:44-PDT From: Liquid Len <Asbed@USC-ECLC.ARPA> Subject: WordStar EGA Patches To: D.Buerger%SCU%PANDA@SUMEX-AIM.ARPA ReSent-To: info-ibmpc@USC-ISIB.ARPA I'm not officially on the net, so I'm sending this to you directly instead of to the bulletin board. I also don't have an EGA, so I'm not sure whether the following patch works. According to the 12/15/85 issue of "PC," you can create a file called SCRIPT with the WordStar document mode. The file should contain the following lines (note the EXTRA blank line in two spots): N WS.COM L N WSEGA.COM E 248 2B A 2A4 JMP 2E0 NOP JMP 302 NOP A 2E0 MOV AH,03 MOV BL,00 INT 10 MOV [0308],CX MOV AX,0003 INT 10 MOV AX,1112 MOV BL,00 INT 10 MOV AX,1200 MOV BL,20 INT 10 MOV CX,0007 JMP 030A MOV AX,0003 INT 10 MOV CX,0000 MOV AH,01 INT 10 RET W Q This setup is for color monitors. If you have a mono system, change the two MOV AX,0003 lines to MOV AX,0007. Then, with a copy of DEBUG (version 2.0 or above), WS.COM, and SCRIPT in the same directory, type DEBUG < SCRIPT. This creates the new version of WordStar, called WSEGA.COM. I hope that this works. If it does, would you please pass it on to the net? Unfortunately, I don't have a patch for a bigger column setup. Best wishes. Vincent Alfieri, Ph.D. (213) 257-9866 ------------------------------ Date: Mon, 14 Jul 86 13:07 AST From: <IUS%DACTH51.BITNET@WISCVM.ARPA> (Eberhard W. Lisse) Subject: Re: TEKTERM, PUSHDIR/POPDIR To: Info-IBMPC@B.ISI.EDU >From: Edward_Vielmetti%UMich-MTS.Mailnet@MIT-MULTICS.ARPA >I have found a Tektronics emulator called TEKTERM. It was developed >at Virginia Tech (anyone listening from there?) and written by >Greg Sherman. It seems to work, but I think I have a damaged copy. >I'd suggest that you get in touch with Va. Tech; I can't point you >to any bulletin boards or anything because I've only seen it on >the board where it was bad. Greg Sherman's user id is LEONARDRG@VTVAX5.BITNET. I have a working copy here and will send it later this week to GKN3M2@IRISHMVS. I am also willing to send it as HEX file (I use HC.ASM which I can also mail) to anybody on ARPA who then can post it to USC. (There is no source, though). I use Kermit v 2.29 and run TEK.EXE from it whenever I need to look at graphics. That works good enough. We have also Coefficient Systems VT100/4010 emulator, which is a bit better but does not work with the german keyboard layout. >From: Brent W Baccala <baccala@USNA.ARPA> >In PC Magazine, Vol. 5 Num. 10 (May 27, 1986), John Friend presents >his PUSHDIR/POPDIR programs ("Programming Utilities: Stacking up >Subdirectories", p. 243). These programs maintain a stack of directory >names, and function as their names imply. They are written in assembly >code for DOS 2.0 and later and feature a default stack holding 6 >directories, but this can be changed. The best part of all is that >they are FREE. > >There are a number of ways to get the goodies. The most preferable is >to download them from PC's Interactive Reader Service at (212)696-0360, >which supports Xmodem protocol. Two BASIC programs to create the COM >files are also provided, in case you can't download the binary and don't >have an assembler. Second, you could type in the source from the >article (Ugg!) and assemble it. Finally (if all else fails), drop me a >line explaining your problem and we'll arrange a solution. Pete Galvin (CC.GALVIN@R20.UTEXAS.EDU) has mentioned he has PC-Outline in his public directory. The ARC file includes PUSHDIR and POPDIR (no sources). FTP it from there. On BITNET I can send it. I use them a lot. el ------------------------------ Date: Tue, 15 Jul 86 15:23:23 EDT From: Brent W Baccala <baccala@USNA.ARPA> To: brackenridge@usc-isib.ARPA Subject: PUSHDIR/POPDIR ReSent-To: info-ibmpc@B.ISI.EDU PUSHDIR.ASM PUSHDIR will push the current drive and working directory onto a stack capable of holding 6 directories (this can be expanded by changing the source and re-compiling). POPDIR will pop the last directory pushed by PUSHDIR. It checks first to see if PUSHDIR has been run, and if not, gives an error message. John Friend, PC Magazine, Vol.5 Num.10, p.243 Brent W Baccala <baccala@USNA.ARPA> 7/15/86 [PUSHDIR.ASM (which contains source for both PUSHDIR and POPDIR) has been added to the library. -wab] ------------------------------ Date: 15 Jul 1986 19:37:25 PDT Subject: GDIR, REMIND, SETTIME, ERROR Added to Library From: Billy <BRACKENRIDGE@B.ISI.EDU> To: info-ibmpc@B.ISI.EDU Last month Robert Lenoil sent the info-ibmpc library several programs. I put the programs in the library and updated program-library.list, but forgot to announce the programs in the digest. Sorry about that. We value the contributions we get. Here are the abstracts from the program library: GDIR.C Gdir is an enhancement over DOS's dir command. It includes file attributes in its display, allows you to perform directories of hidden and system files, and also allows you to recursively check all of a directory's subdirectories as well. Robert S. Lenoil <LENOIL@XX.LCS.MIT.EDU> 6/12/86 REMIND.ASM is a program that can pop up a small reminder on line 25 of your screen at a given time. It displays messages via the ROM BIOS, and thus will work on any IBM BIOS-compatible machine with any display hardware in any graphics mode. The syntax is: REMIND time message, where time is one or two hour digits followed by a colon and two minute digits. REMIND can also be entered without any arguments, in which case the pending, or if none, the previous message is shown. Program requires SETTIME.ASM. Robert S. Lenoil <LENOIL@XX.LCS.MIT.EDU> 6/12/86 SETTIME.ASM is a program to set the BIOS timer to the correct time of day. It is useful mostly for programs like REMIND.COM, which require the BIOS timer to accurately reflect the time, even though DOS may be getting the time directly from a clock card, via the clock.sys device driver. The syntax is: SETTIME. SETTIME gets the time from DOS, and writes it to the BIOS timer. Robert S. Lenoil <LENOIL@XX.LCS.MIT.EDU> 6/12/86 From <info-ibmpc.eel>eelcode.list ERROR.MAC The macros place the error message output by the compiler or assembler in a one-line window at the top of the screen, and then position the cursor to the offending line in the source-code window. It seems that these macros will not work with Epsilon 3.x, as the macros use the PROMPT command from Epsilon 2.x, and this doesn't seem to be present in Epsilon 3.x (probably an oversight by Lugaru). Robert S. Lenoil <LENOIL@XX.LCS.MIT.EDU> 6/12/86 ------------------------------ Date: Wed 16 Jul 86 15:41:15-EDT From: Frank da Cruz <SY.FDC@CU20B.COLUMBIA.EDU> Subject: SCRNCOLR.ASM Added to Library To: info-ibmpc@B.ISI.EDU Here's a version of SCRNCOLR.ASM that, unlike the version in your library, will assemble and run. SCRNCOLR - Change display screen from white to color attrib. Adapted from "SCREEN.ASM" by H. Fischer - HFischer@eclb 12/83 SCREEN was Typed in by Glass - gjg@cmu-cs-cad from Byte Nov. 1983 [The new SCRNCOLR.ASM has been added to the library. -rag] ------------------------------ Date: Sun, 13 Jul 1986 17:34 EDT From: James H. Coombs <JAZBO%BROWNVM.BITNET@WISCVM.ARPA> Subject: LEX && YACC To: <INFO-IBMPC@USC-ISIB.ARPA> There are at least two versions of YACC for the PC and one of LEX. You can get them from the IEEE bulletin board in Rhode Island (401) 849-0529. Someone in Texas is also offering YACC and LEX for $25 each (I have vague memories of "Austin" and "Code Works"--there is a note on BIX in the LATTICE/UTILITIES conference). Vernon Buerg also has a version of YACC--(415) 994-2944. Actually, I am surprised that these are not in the INFO-IBMPC library. Since IEEE has source for YACC and LEX, they could be submitted. I would be happy to do it unless someone else has two or three free hours before I do. (There are a lot of files involved unless ARCs are acceptable). Some YACC/LEX combination was written for DeSmet C. There has been some brief discussion of this on BOSS, and I believe that someone has a version for Lattice (BOSS is (201) 568-7293--CLANG conference). --Jim Acknowledge-To: James H. Coombs <JAZBO@BROWNVM> ------------------------------ Date: Sun, 13 Jul 1986 17:57 EDT From: James H. Coombs <JAZBO%BROWNVM.BITNET@WISCVM.ARPA> Subject: Interactive EasyFlow To: <INFO-IBMPC@USC-ISIB.ARPA> HavenTree has an "Interactive EasyFlow" for developing flow charts and organizational charts. Works with CGA and Hercules. Features 40 predefined shapes and automatic line routing. Facilities for drawing your own lines makes the program useful for data modelling as well (although one would not want to say that the program actually supports data modelling). The program works as advertised, and I have seen no signs of bugs. The manual could use some work, but all of the necessary information is there. I cannot make a direct comparison with MS Chart and the like, but EasyFlow is only about $150 and probably better suited for flow charts. If anyone knows of anything affordable that supports data modelling, I would like to hear about it. --Jim Acknowledge-To: James H. Coombs <JAZBO@BROWNVM> [You should also check out the block diagram editor "Boxes & Arrows", available from Inner Loop Software 213-822-2800. But since this is my company, I can't claim to be unbiased. -rag] ------------------------------ Date: Mon, 14 Jul 86 11:21:23 cet To: INFO-IBMPC@USC-ISIB.ARPA From: FAFKH%NOBERGEN.BITNET@WISCVM.ARPA Subject: Simula Simula will be available for MS-DOS October 1 1986, see Simula Newsletter no 2, May 1986. The price will be Nkr. 10.000, Nkr. 5.000 to universities. Simula will need a 8087 or 80287. Information about the configuration that is needed to run the system is not given, but it will run on a "large" system. Contact: Simula a.s. Dag Hammarskjoldsvei 35 N-0585 Oslo 5 Norway Tel. +47 2 152900, telex 77458 simx-n Knut. ------------------------------ Date: Tue, 15 Jul 86 04:31:34 PDT From: Ya'akov_Miles%UBC.MAILNET@MIT-MULTICS.ARPA To: info-ibmpc@USC-ISIB.ARPA Subject: BASIC Compiler File SMALLERR.OBJ If you explicitly link to the SMALLERR.OBJ file, you will get a smaller compiled program, without the ASCII text of the error messages... ------------------------------ Date: Tue, 15 Jul 86 17:34:53 PDT From: gts%ucbviolet@BERKELEY.EDU (Greg Small) To: info-ibmpc@b.isi.edu Subject: PC Net 1.0 vs PC-DOS 3.2 The PC Network 1.0 server can be run on DOS 3.2. Just be sure to use the SHARE.EXE from DOS 3.2 not the one that came with PC Network 1.0. SHARE.EXE and NETBIOS.COM abort if they are used on the "wrong" version of DOS. Greg Small (415)642-5979 Personal Computer Networking & Communications gts@ucbopal.Berkeley.EDU 214 Evans Hall CFC ucbvax!ucbjade!ucbopal!gts University of California SPGGTS@UCBCMSA.BITNET Berkeley, Ca 94720 ------------------------------ Date: Mon, 14 Jul 86 22:27:18 PDT From: WU%TOM@ames-io.ARPA Subject: Hard Disk, Plotter Queries To: IOinfo-ibmpc-request@isib 1. I recently purchased a Seagate 4051 40 MB Hard Disk for an IBM AT Clone. Unfortunately the Western Digital Controller only recognizes 854 of 977 tracks (Type 11 Disk) for a total of 36 MB. Is there anyway to use the rest of this disk? 2. The Epson HI-80 Plotter was highly recommended in a recent PC Magazine. The disposable pens supplied by Epson seem to last for about 10 plots. Epson does not supply refillable pens. Does anyone know of any third party suppliers of pens for the HI-80? Send replies to wu@ames-aero. Thanks. Alex. ------------------------------ Date: Tue 15 Jul 86 11:10:22-CDT From: Pete Galvin <CC.GALVIN@R20.UTEXAS.EDU> Subject: ATACK AT Motherboard Query To: info-ibmpc@B.ISI.EDU Has anyone tried to build an AT by starting from scratch with the ATACK motherboard? for $99, you buy the board and supply all your own chips and support hardware. I was wondering how compatible the board was (wouldn't it be a drag if you built the whole thing and found it couldn't run your favorite program?) --Pete ------------------------------ Date: Tue, 15 Jul 86 12:17 EDT From: Manasseh Katz <MKATZ%UMDD.BITNET@WISCVM.ARPA> Subject: Problems with New IBM PC/XT To: INFO-IBMPC@USC-ISIB.ARPA I have had problems with Framework, and now with DESQview on a new IBM PC/XT. In Framework, the Scroll Lock key does not function - it works fine in SideKick (I can't think of anything else that uses it). In DESQview, the ALT key does not work, which makes DESQview almost useless. The ALT key does work for calling up SideKick (Ctrl-Alt) and within a DESQview application (e.g. ALT+numeric keypad to produce graphics characters in any application), but it doesn't work in DESQview to call up the DESQview menu. Both of these problems occur on two new XTs. DESQview works fine on an AT and on an older XT. Framework works fine on older XTs. The problem is not the keyboard - the new XTs came with old keyboards, and just in case IBM had changed the old keyboard too, I tried a keyboard from an older XT and the problem didn't disappear. I am using DOS 2.1. If someone out there has the same problem or doesn't have the problem or has a solution to the problem, please send me mail. Manasseh Katz MKATZ@UMDD.BITNET ------------------------------ Date: Tue, 15 Jul 86 16:23:32 edt From: Mike Ciaraldi <ciaraldi@rochester.arpa> To: info-ibmpc@usc-isib.arpa Subject: Display-Write 3 File Recovery A coworker using Display-Write 3 has developed a problem with an important document, and I am trying to help recover it. She was editing the document (about 100 pages), making a lot of changes to it. Then she exited DW3 without saving. Now when she tries to edit the document with DW3, it gives the message "Disk Error" and won't let her open the document. DW3 does not give any more detail than that, and there is no DOS message. I have checked the disk--there are no bad sectors, crossed links, or anything like that. I can copy the file with no problems. But DW3 will not accept the file, with either the Revise command from the main menu or the Recover command from the Utilities menu. She realizes that her latest edits are gone, but is there any way to recover the old version of the file? Even part of it would be useful, to minimize retyping. Things I already tried: Looking at the file with Ultra-Utilities. It just looks like garbage. Considering DW3 stores its text in EBCDIC, this is not surprising. Converting the file from EBCDIC to ASCII. This produces sections of the original text, interspersed with garbage (mainly control characters). However, some sections of text are repeated, and the various sections that appear are not in the same order as they appear in the original text. Our local IBM office's answer was, "Use your backup copy." Naturally, she had never made one. So, I am hoping someone out in netland can tell me how to recover the document from this file, how to trick DW3 into taking it, or whether there are some programs that might help recover it. Info on the internal structure of DW3 files might help, too. Thanks. Mike Ciaraldi ciaraldi@rochester seismo!rochester!ciaraldi ------------------------------ Date: Wed, 16 Jul 86 11:33 EST To:info-ibmpc@usc-isib.arpa From: GKN3M2%IRISHMVS.BITNET@WISCVM.ARPA Subject: EMS Board Query Are there any clone owners who have successfully installed EMS boards in their PC? What about unsuccessfully? Which are the boards to avoid? Does MS Windows use EMS memory in any way other than just a big RAMdisk? As the release date for DOS 5.0 draws nearer, those of us without 80286 chips may want to install accelerator cards. What about compatibility with any EMS/EEMS boards? Thanks in advance. Evan Bauman University of Notre Dame gkn3m2@irishmvs (BITNET) gkn3m2%irishmvs.bitnet@wiscvm.wisc.edu (ARPANET) ------------------------------ Date: Wed, 16 Jul 86 14:01:27 PDT From: hunt%pavepaws@BERKELEY.EDU (Jim Hunt) To: zen!B.ISI.EDU!Info-IBMPC Subject: Special Multi-Function Card Wanted I have a clone (eagle PC2, I have the roms) with 512 but only three slots and have reached the point where I can no linger put off adding a hard disk. Xebec's floppy/hard controller will save me a slot. Leaving the other two for video and modem. What I am looking for is a way to get a clock calendar in the same effort. Does anyone make a dual floppy/dual hard controller with a CLOCK too, or how about a video board with a clock? Modem? jim hunt seismo!ucbvax!pavepaws!hunt ------------------------------ End of Info-IBMPC Digest ************************ -------