Info-IBMPC@C.ISI.EDU (Info-IBMPC Digest) (04/21/87)
Info-IBMPC Digest Monday, 20 April 1987 Volume 6 : Issue 29 This Week's Editor: Billy Brackenridge Today's Topics: CMU pcip via ARPANET Followup to DCA/DIA information request. Clone Test from Computer Shopper verifies BIOS Turbo BATCH Keystroke Mods Two Color Monitors at Once Offer to Send FreEmacs CMU MIDI software Microsoft DIAL DOS Copy OK Hard Disk Bad Brief Reviews of 3 Text Data Base Management Systems DIAL.SRC Posting of DIAL Program Another Great Product Screwed up by Programming Methodology Logitech Mouse and Windows Ill Behaved Programs can Trash File System SPL Shareware Announced Today's Queries: Does TDEBUG support Turbo Pascal V3.02a? Accelerating the Leading Edge Honeywell VIP7300 and/or 7800 Terminal Emulators DOS 2.1 and Disk Space Display Write 4 with Laser Jet Plus Control-Shift-Alt-key Combinations 8087-80287 Problem Disk Recovery Programs Graphics toolbox for TurboPascal wanted INFO-IBMPC BBS Phone Numbers: (213)827-2635 (213)827-2515 ---------------------------------------------------------------------- Date: Thu, 16 Apr 87 21:56:40 edt From: ddp#@andrew.cmu.edu (Drew Daniel Perkins) Subject: CMU pcip via ARPANET Since someone forwarded my mail about CMU PC/IP on bitnet, people may as well know how to get it from the arpanet too... The TCP/IP package for PC's better known as PCIP was developed at MIT using a cross compiler on a UNIX VAX. In order to make development using this package easier, it has been "ported" to the Microsoft C compiler and Microsoft Macro assembler. You must have version 3.0 or higher of both the compiler and the assembler. Also you must have the Microsoft Make program, Librarian and linker that are distributed with the compiler and assembler. To get the CMU Microsoft C version of the PCIP package from the arpanet, connect to host "te.cc.cmu.edu" with FTP (no quotes when you really type it). This machine is a TOPS20 system. Login in as user "anonymous", password "guest". Next, use the "cd" command to change your working directory to "pk:<pcip>". Now if you do a "dir" command you will get a listing of all the necessary files. First, "get" the files "readme" and "install.bat" in netascii mode. The rest of the files must be retrieved in binary/octet mode. On a UNIX system use the command "tenex" to tell TOPS20 to use a local byte size of 8 bits. Now retrieve the files "tarread.exe", "root.tar", "include.tar", "srcdev.tar", "srclib.tar" and "srccmd.tar". The file "doc.tar" is also available if you want the scribe documentation. Alternatively, you can now retrieve the files via anonymous FTP from host "lancaster.andrew.cmu.edu". This host may not yet be in your host tables, but should be resolvable via the domain name system. It's IP address is "128.2.13.21". This machine is a 4.2 bsd UNIX system. After you log in, use the "cd" command to change your working directory to "pub". You will have to retrieve the same set of files as above. Once you have these on your local machine, use TFTP or some other file transfer program to get them to your PC. Put the files under a subdirectory such as c:\pcip. Make sure you do the transfers in the proper mode (octet or ascii, as above). The file "readme" explains what you have, and how to proceed farther. The program tarread.exe is a very small tar file reading program that I wrote. It has very few features, but it serves the purpose here. [Addendum from a later message follows. -wab] Due to problems people are having with the arpanet, I have added "compressed" versions of the 4 largest .tar files to my distribution. If you anonymous ftp the distribution from lancaster.andrew.cmu.edu there are copies of include, srccmd, srclib and doc with both a .tar extension and a .tar.Z. The .tar.Z files are about 1/3 the size of the .tar files, so ftp'ing them should work much better. Of course if you get the .tar.Z files you will first have to convert them to .tar files using UNIX uncompress. In either case, you don't need both versions of the file, so don't use the FTP mget command! Drew Perkins arpanet: drew.perkins@andrew.cmu.edu phone: (412) 268-6628 US mail: Drew D. Perkins Carnegie-Mellon University 4910 Forbes Ave. Pittsburgh, PA 15213 ------------------------------ Date: 16 April 87 20:37-CST From: AYAC071%UTA3081.BITNET@wiscvm.wisc.edu Subject: Followup to DCA/DIA information request. My thanks to those who responded to my request for references to IBM's Document Content Architecture and Document Interchange Architecture. The list is posted below. Thanks especially to Jim Coomb and Mary McClure of Brown University (congratulations, Amy), and to Gerry Key of nocs.mil. The below are all IBM documents, & should be available through the local IBM office. Order # Title =========== ========================================================== GC23-0765 Office Information Architectures: Concepts SC23-0757-1 Document Content Architecture: Final-Form-Text Reference SC23-0758-0 Document Content Architecture: Revisable-Form-Text Ref. SC23-0764-1 Document Interchange Architecture: Interchange Document Profile Reference SC23-0763-1 Document Interchange Architecture: Transaction Programmer's Guide SC23-0781-0 Document Interchange Architecture: Technical Reference Bill Douglass AYAC071%UTA3081.BITNET@wiscvm.wisc.edu ------------------------------ Date: 17 Apr 87 08:49:00 EDT From: "V703::S_DANIELS" <s_daniels%v703.decnet@nusc> Subject: Clone Test from Computer Shopper verifies BIOS Dan Ishikawa asked for info on a program that "performed a comprehensive test of system BIOS." By that I assume he means checks that the BIOS performs as advertised, and also works like that on a true-blue PC. Well, the latest issue of COMPUTER SHOPPER (May) had an article called ":Clone Test" that does just that. The article describes the tests done on the BIOS, but no code is presented. Top get the actual program, you must send $10 to the publishers. I don't have the address handy. You can probably pick up a copy of CS at any large magazine retailer - we have a "Newstand": in our mall that carries it. Hope this helps. SCOTT ------------------------------ Date: 17 Apr 87 08:55:00 EDT From: "V703::S_DANIELS" <s_daniels%v703.decnet@nusc> Subject: Turbo BATCH Keystroke Mods Keith Redwill asked for info a making Turbo Pascal respond to a file of keystroke commands, much like any other DOS program (e.g EDLIN, DEBUG) would, so that he could type: TURBO < keystrok.txt and it would go from there. I have had good luck with the programs FAKEY and KEYFAKE. The former is one of the PC Magazine utilities. The latter is another pd program. Both do the same: "preload" the ks buffer and feed these ks to the application program one at a time. To make Turbo load up, answer the "Load error messages Y/N", and compile a program I would make a batch file for KEYFAKE like this: : TP.BAT keyfake "Y" 13 "C" %1 13 turbo Here the stuff in quotes is the text you would normally type. the "13" is deciaml for carriage return. "%1" gets your batch parameter. The file you want to work on. To run it - type - 'TP test' and it will load and compile TEST.PAS. ANOTHER CHOICE: Turbo Extender from Turbo Power tools has a mod to TP to make it respond to batch files. I haven't played with it much but it works, at least using TE's commands. The TE has a sort of 'make' facility for use with Turbo. Quite useful, for large programs and I recommend it. The other stuff is available on pd bbs's - try ours (203) 886-5265.LUCK.SCOTT ------------------------------ Date: Fri, 17 Apr 87 09:09:08 EDT From: Russell Nelson <bh01%CLUTX.BITNET@wiscvm.wisc.edu> Subject: Two Color Monitors at Once You can indeed install two egas at once, however most egas do not let you disable the rom. The TECMAR has a soft switch on the egas IO address that lets you use the same bios for both egas. The EGA wonder is an ordinary EGA that will do interlace on a RGB monitor. It won't solve your problem. I think that you could install a CGA and an EGA. You would have to write directly to the EGA's registers to get it to act solely like an EGA. The bios interferes with this because it doesn't support multiple monitors very well/at all. -russ ------------------------------ Date: Fri, 17 Apr 87 09:45:52 EDT From: Russell Nelson <bh01%CLUTX.BITNET@wiscvm.wisc.edu> To: rogers@uiucvmd.bitnet Subject: Offer to Send FreEmacs I assume that you're talking about an Emacs for the IBM-PC. Why not take a look at Freemacs? It's a lot smaller and faster than any other Emacs for the IBM-PC. You can get it from Simtel20 or I will mail you a uuencoded arc file. -russ ------------------------------ Date: 17 Apr 87 12:13 PST From: JMH%SLACVM.BITNET@wiscvm.wisc.edu To: INFO-IBMPC@USC-ISIB.ARPA Subject: CMU MIDI software In response to my request recently for the CMU Midi driver I got the following response; The Carnegie Mellon package is still available for $20 (payable to Carnegie Mellon) to: Roger Dannenberg Center for Art and Technology Carnegie Mellon Pittsburgh, PA 15213 Currently, only a Lattice C version is available, but MS support will be done by September. James M. Hodgers Stanford Linear Accelerator Center ------------------------------ Date: Fri, 17-Apr-87 19:48:30 PDT From: bcsaic!asymet!fred@june.cs.washington.edu (Fred Wamsley) Subject: Microsoft DIAL My company subscribes to DIAL and has gotten useful information through it. I question whether it's cost-effective for an individual, though. The quality of information on DIAL is fairly good, at least for Windows. You can browse through a database of selected problem reports from other developers along with their resolutions. There is a much better signal to noise ratio than on some place like Compuserve. A PC or clone is required to run the local end of the DIAL software. This provides a front end that automates submitting Technical Assistance Requests (TARs), provides ring menus for navigating through DIAL, and crashes your machine occasionally. We have found the quality of technical support through DIAL to vary wildly, so I won't try to comment on it. If you're developing for Windows, Microsoft also delivers on line support through the GEnie network. Many developers find GEnie easier to use and more helpful. To summarize: if you're writing software for money, buy every resource you can get, including DIAL. If not, spend the cost of a DIAL subscription on a hard disk. [Info-IBMPC has tried to get access to a DIAL account, but we have no money and it is strictly pay as you go. As I understand the shrink wrap non disclosure agreement (A new legal concept!!) you must keep confidential the information gleaned from DIAL. As always INFO-IBMPC shields our sources. If anybody picks up any technical gems from DIAL or any other sources we will publish. (subject to our usual good taste etc.) -wab] ------------------------------ Date: Sat 18 Apr 87 15:13:27-PDT From: JOHN R. THOMPSON <WOOLFORD.THOMPSON@BIONET-20.ARPA> Subject: DOS Copy OK Hard Disk Bad DOS copy problem solved. > I discovered a curious problem with the DOS 3.2 COPY command. > Several attempts to copy a 1.3 megabyte file failed to produce a > verifiable copy (judged by using the DOS COMP command). The file was > correctly copied using PCTOOLS. Is COPY incapable of handling large > files???? My problems with DOS copy were apparently related to a flaky hard disk. The Disk on which the problem was encountered went belly up several days after the above report; would not boot or accept a low level reformat. After installing the replacement disk the same 1.3 MB text file was copied correctly by DOS COPY. However, I still do not understand why the verify option of copy did not report a problem with the copied file on the flaky disk. The Case of the Vanishing Seagate Warranty: The hard disk that died was a 3 month old Seagate ST238 (attached to a SMS OMTI 5527 controller). The controller was vindicated when a Western Digital controller also failed to format the disk. (Anyone know if it is possible for a controller to physically damage a disk?) The company I purchased the drive from had gone under since the purchase. To their credit, Seagate honored the warranty for a small fee ($35). However, Seagate will only honor the warranty for one year from the date of manufacture (stamped on the drive) not from the date of purchase. Not too cool in my opinion since to my knowledge Seagate does not sell direct to the public and we certainly have no control over how long the things sit on the shelf at the suppliers. Furthermore, Seagate swaps out a reconditioned drive for a bad returned drive (the swap took 7 days through the US mail). The catch is the replacement drive is warranted for a mere 90 days! All in all my 1 year warranty has now shrunk to 6 months from date of purchase. Incidentally, I know of three people that own this drive/controller combination. Two of those three have had failed disks after no more than three months operation. Not a statistical sample but certainly not a testimonial either. ------------------------------ Subject: Brief Reviews of 3 Text Data Base Management Systems Date: Sat, 18 Apr 87 22:08:51 -0700 From: Rob Kling <kling%icse.UCI.EDU@ICSE.UCI.EDU> Comments Re. Text databases: Ask Sam 3, Notebook II (v 2.3) and Dayflo 1.3 Rob Kling Computing, Organizations, POlicy & Society (CORPS) Department of Information and Computer Science University of California - Irvine 4-17-87 I'm responding to an inquiry and several brief comments about text databases (TDBMS) for IBM-PC's. I use several different TDBMS for different applications including maintaining annotated bibliographies, documenting statistical data files for research purposes, keeping to-do lists, managing mailing lists, managing a phone book, keeping a restaurant file, keeping notes about research sites, maintaining narrative commentaries about the status of a set of related projects, etc. I find TDBMS to be an essential tool in my daily work, but find the state of the art products not meeting my preferences. Even so, I use some TDBMS because they are useful in different ways and provide tremendous advantages over the alternatives: various paper systems, text editors, outliners, and DBMS with short fields (80-255 characters and text facilities such as word wrap, margin settings, block copy and mode, etc.). I use three different TDBMS: Notebook, Dayflo, and AskSam. Each of these meets some minimal requirements: 1. Allows fields to include lots of text (often up to 28K) 2. Gracefully handles word wraps at the margins when text is entered. 3. Includes a text editor with abilities to copy, move and delete blocks of text. 4. Abilities to search bodies of text easily for particular words. 5. Some kind of report writer integrated into the package. These is no common core of requirements for specifying a TDBMS which all systems currently meet. This area is less well understood than, say, spreadsheets, text editors, and other kinds of databases. I suspect that each TDBMS is designed with some narrow range of applications and users in mind. While each is subtly advertised as universal, none is. Period. I would like to consolidate all my active text databases into one system. Each of these three TDBMS could serve as a medium for all my text databases, but none would do all the jobs very well. And none of these can serve for some other needs in organizing and coding interviews and qualitative field data - a kind of SPSS for ethnographic research. Each of these 3 systems is powerful in different ways and not one is ideal. None of these databases is really good for taking a raw body of notes and slowly organizing them into some structure. (A good outliner, like PC-Outline may be better for that kind of task.) Notebook II is a menu driven system which requires that each record have the same field names. The fields and records can be very long (about 28K max ). It has facilities for easily adding fields to records & restructuring records. Its searching is quite fast and its report writer is OK. It's relatively fast. It is menu driven and does not include any macro capability. For me, its biggest limitation is that all fields are of one type: ASCII characters. Dates and numbers are interpreted as ASCII strings. Dates must be represented as strings like 87/04/15 to have date comparisons work properly. Numbers must all be represented as the same length strings. $12 < $9, but $12 > $09. Notebook has evolved through several versions (to version 2.3 ) with increasing refinement at each stage. Notebook handles pages of text rather gracefully, searches them rapidly and stores them concisely, and I find it to be the best of the lot for annotated bibliographies. I have three annotated bibliographies in Notebook, and two of them have over 600 entries apiece. Notebook handles these the best of all three databases. It also has some features, such as the ability to compile a list of "keywords" -- all the words used in any field within a file, or a list of all the words used in a file, including counts of each word. In contrast, AskSam is clumsy in handling pages of text in a field that run longer than 20 lines. The latest version (v 3) has a "document mode" which lets one string together a set of 20 line shorter segments of data, but it does not automatically reformat text across these segment boundaries when new text is inserted in the middle of existing text blocks. Dayflo is relatively slow on an 8088 based machine, and its file storage is more space consuming. (I'm told that it zips along nicely on an 80286 based machine.) I find Notebook's one type of field to be it's greatest disadvantage since it makes dealing with dates and numbers cumbersome. I also dislike its screen layout which forces each field to take at least one line. I would sometimes like to place several short fields, such as a name, city and date, one line. In both AskSam and Dayflo, a record can be any arbitrary collection of fields. This makes it easy to include unusual records in any database - such as a description of the database, where it is archived, etc. as a special record. In practice, many records will have the same structure, but they need not. For example, a bibliographic entry for an "article" usually includes a journal/source, volume, issue #, etc. A record for a book doesn't include these fields, but may include the city/country of publication. In Notebook, records with such different sets of fields would have to be represented as different files (or one would create a "superrecord" which had all the fields.) In Asksam and Dayflo, records representing books and articles can easily co-exist in the same database, along with records which have many different structures (e.g., for films, etc.). Dayflo and AskSam allow a user to place fields (and data) in any locations on the screen. They also support multiple field types, but also allow the types to be overridden. (Actually data is typed in Dayflo and can be treated as if it were typed in ASkSam so that data and numerical arithmetic work well.) Dayflo is the most elaborate of these systems and has the slickest screen display. It is very polished. The product is menu driven, but also includes macro facilities to simplify certain repetitive tasks. Dayflo is also unique in allowing records to be placed on separate "stacks." There is no restriction about the format of records on any stack. Stacks can be used to help organize records in meaningful sets. For example, in a bibliographic database, one stack might contain records of books and articles to be gotten from the library while another stack might contain books loaned to friends. This kind of information can be handled in any DBMS by adding a status field to indicate books on loan, etc. and using the search routines or report generator to pull out a list of records that meet some status (e.g., get them from the library). But Dayflo's stacks offer a novel and useful different way to segment information and keep order. It is also easy to browse through the separate stacks. Initially, Dayflo is the most daunting system since it comes with two large manuals and several smaller booklets. But it's a well designed system. Its tutorials, menuing and help facilities ease learning. AskSam does support some powerful reporting and record restructruing capabilities. It includes a powerful, but somewhat cryptic, text programming language. Of all these three systems, it is the only one where one can actually write a program to alter numbers and dates with arithmetic operations: e.g., add 30 days to a set of dates in the database, add 10% to a set of prices. AskSam is command driven and is very crude in many areas. Programs that generate reports or alter data are also records. This is a powerful concept, loosely akin to having LISP functions be LISP code. But these program records are mixed in with all the other data records in a data file and one must be careful to structure reports so that they print out genuine data records and don't sweep up other programs in record form when they scan the data file. AskSam's "help" system is the worst that I have ever encountered in a system which claimed to support "help" since it throws the user out of his place in the file he is working on and load a new "help file" in its place (See below). It sets the concept of interactive "help" back 15 years. AskSam has been the subject of several overlay enthusiastic reviews. I bought it last December, struggled with it for several weeks, and finally sent a long unsolicited letter to its designers. I received only a cursory acknowledgement of my letter, so I have no reason to believe that any of the problems I've identified will be fixed in the near future. I've excerpted from my letter below, as a way of indicating AskSam's several Achille's heels and providing further contrasts between these products. I have been a beta tester for Notebook-II and a customer for Dayflo and Seaside Software (Asksam). I have not been monogamous, and use each of these products for some application. If speed and space were not an issue for me, I would move to Dayflo. Which may mean that Dayflo indicates that graceful and robust TDBMS are difficult to squeeze into 8088 machines with 360K diskettes and many other programs sharing the hard disk. (In Dayflo's case, the size of the programs is not as much of a problem for me as the rate at which the databases swell. The smallest database starts out at 100K and can swell to 500K with only 200 records. Dayflo includes facilities for backing such databases across multiple floppies. But I have a lot of different little databases. Dayflo could absorb them all into one humongous "superbase," but that is not how I choose to work.) Asksam is a good bet if you have very short text fields (say 10 lines or less) and want to have good numerical and date arithmetic. You also have to want to program to get meaningful reports. There is a programming language which generates reports, but no report writer for selecting parts of records. Notebook is the best bet for someone who needs flexible manipulation for lots of text, such as annotated bibliographies. Both Notebook and AskSam will operate with floppy based machines and save relatively concise files (staring perhaps around 2KB). In the end ... if someone has found a "universal" TDBMS, please contact me ASAP at UC-Irvine. :-) Dayflo is sold by Dayflo Software (17701 Mitchell Ave, N. Irvine California 92714. 714-474-1364). Dayflo 1.3 lists for about $700, including the report generator (Reportflo) and some very useful sample applications. It is sold to universities for about $350. Dayflo just came out with a somewhat simplified variant, called Tracker, which is being offered for $100 with a variety of sample applications in the next month or two, before the price jumps up to about $140. Notebook II v 2.3 is sold by Pro/Tem (814 Tolman Drive, Stanford Ca 94305). Notebook lists for about $180, but I have seen it for well under $100 at a nearby discount software store. It can be site licensed to universities at a very low price per unit. AskSam 3 is sold by Seaside Software in Perry, Florida. It sells for about $200. =========================================================== (Excerpts from) Comments Re. Ask Sam 3 (with reference to Dayflo 1.3 & Notebook II v2.3) Rob Kling Department of Information and Computer Science University of California - Irvine January 1987 A. My environment My environment: I have a Leading Edge Model D PC with 640K and a 20MB hard disk running MSDOS 2.11. B. Some background For the last three years I've used Notebook (I & II) to build files about phone lists, books and articles (with annotations and reading notes that run up to 75 lines), (and frivolously, restaurants that I frequent), etc. Most of these databases have several hundred records each. (The books and articles databases are the largest, with about 600 records and growing monthly.) I have been a beta tester for Notebook for over 2 years. I have found Notebook to be a versatile, useful, but somewhat unexciting product. Even so, I keep on returning to Notebook for new applications because its tradeoffs seem to work well in favor of the overall ease of designing and using applications. I have been discouraged by some of Notebook's limitations, especially the weak handling of dates/numbers & limited text editor. One key application is a tickler file. I have moved it from Notebook to 1-2-3 to Enable's DBMS to Reflex (in early 1986). I am now trying that application and two others, in AskSam. Since last July, I've been using Dayflo in addition to Notebook and Reflex. I have built several databases including a large contact list (about 1500 names, addresses, phone numbers, interests), an equipment inventory for my research projects, tickler files of people to contact and upcoming commitments. I basically like Dayflo, and found that it has record structures which are substantially more flexible than AskSam's, a workably clean interface and good reporting capabilities. I have found Dayflo and Notebook fairly easy to teach temporary part time student assistants how to use. But I have found Dayflo to be generally sluggish since it goes to disk for many of its overlays and for moving around records. It is unacceptably slow in some key operations (e.g., on an 8088 based PC, it can take an hour to print 1500 records to disk; it can take 3 hours to reindex that database). Dayflo requires the use a hard disk machine. It builds large files, making archiving more time consuming and space consuming than it should be. Overall, I'm least happy with Dayflo's sluggish interaction. [I have found AskSam to be very quick with the small databases that I've built so far, but I wonder how it will work with 2000-3000 names and addresses, when that database grows. I have found Notebook to be much faster than Dayflo in overall performance. Oddly, Vincent Puglia criticizes Notebook for sluggishness in his Fall 1986 PC-Magazine review and doesn't mention Dayflo's speed problems.] I have also been using Reflex for financial reporting, a critical tickler file, managing a special contact list, and for managing the flow of articles for some journals which I edit. I have tried these various text DBMS because they offered special attractive features. However, each has limitations which make it problematic as a general database for my purposes. I hoped that AskSam could serve as a single DBMS for all (or at least most of my text applications). That is what Puglia's review and AskSam's promotional material lead me to believe. D. I have found AskSam to be flexible in some ways and potentially very powerful. I like the fact that AskSam handles dates and numbers with their own logics, as well as text. The reporting capabilities are impressive and the ability to print multiple records per report line appears very useful. Some of the power comes through a programming language which I find OK, but which would be very problematic to teach to my short term assistants. However, I find AskSam needlessly rigid in handling long text fields and clumsy in small and unexpected ways. Many rave reviews mislead about AskSam since they often miss some glaring weaknesses. AskSam is a TDBMS with a certain style: 1. A few menus; 2. A complex command language which can't be avoided for almost any kind of formatted retrieval; 3. A control structure which is made powerful by several modes; 4. A overall environment which is made complicated by having to track the current modes in force (Document/Record, Stream/Image, Sam/ASCII). One has to learn about and manage the interactions between modes. If more flexibility in later releases comes with more modes, then mentally managing the interactions between modes could become another club foot. E. AskSam would be superb if I were to use it primarily as a 3X5 card system with keyword retrieval. I suspect that many reviewers use it for collecting random short notes, contacts, leads, etc. It would be superb for that kind of application -- as long as the notes are less than 20 lines (see below). The freeform text capability is fine as long as the notes are short. However, all but the most trivial reports have to be rewritten to change the query criteria. That means I have to teach any assistant how to edit a somewhat jarring query language without inadvertently clobbering other parts of a report format. I would NOT call AskSam "easy to use" as an overall package, even though parts of it are very simple. (It's the parts of AskSam that aren't simple which cause the problems.) AskSam appears adequate for my trial databases to document data files and list casual contacts. It is a bit weaker as a tickler file manager. But I have the MOST trouble envisaging AskSam as a TDBMS to be used for my annotated lists of books and articles. F. LONG TEXT FIELDS An ideal TDBMS would allow the user to enter unlimited amounts of text in any field and reformat and reorganize the text very flexibly. It would mix a fairly powerful editor/formatter with a good retrieval capabilities. Notebook allows a record to be up to 28KB long. Any field can be as long as a whole record - 28KB. For my purposes, this is long enough. Most of my comments about books and articles are only a few lines, but a few of them run to 6K. AsKSam's field limit of 20 lines (1.6KB) is short indeed. I spoke about AskSam's field size limit of 20 lines with one of the technical support staff. He suggested using multiple variants of a field in Document mode and printing with {ALL} enabled. This strategy sounds plausible as an off the cuff suggestion, but has two flaws, given AskSam's other limitations. First and foremost, AskSam one cannot easily insert new text or move existing blocks of text around within a document since the text will not move across record boundaries when needed to accommodate a new long block. I sometimes want to add additional comments in the middle of an articles description that I've already written. The user has to manage the target space into which new blocks of text are inserted or moved to insure that there is sufficient room in a record. This can mean having to move parts of existing text into adjacent records within the document. This makes document mode a "hack," and a clumsy hack for a true TDBMS. The second dilemma with this strategy is that the multiple field mode {ALL} seems to print field values at the left margin, REGARDLESS of the margin's settings. This makes an outdented report of this form impossible with {ALL} on: Jones, Malcolm. TDBMS Design in 10 Easy Lessons. Yoho Press. [description......................................... ............................................................ ......................30 line description.................... ............................................................ ............] Smith, Audrey. The Cold War in Context. Ypslanti Press. [description .............................. ............................................................ ................40 line description .......................... ................................................ ........ ............................................................ ............................................................ .......] If the descriptions run longer than 20 lines, they will be continue at the left margin. Having to specify LONG fields in reports is also a hack. The user of a TDBMS should NOT have to think about the length of text fields and should not have to manage field types/lengths when writing a report. The TDBMS should manage that work. I can understand why you implemented LONG fields as a special data type after you allowed fields to be acceptable without a terminating ] if they were one line long. This problem is more manageable for me than the rigidity or moving/inserting blocks of text into records in document mode. I find that AskSam's clumsiness and rigidities in manipulating long text fields are its most crippling aspects for my applications. I think that this is a deep problem with AskSam's architecture based on 20 line records. Document mode as a set of linked records is a hack if fields can't run across records and if the text editor won't slide text across record boundaries to accommodate insertions and deletions of paragraph length textual items when they won't fit into an existing record. The examples in your supplementary manual to version 3.0 show that you think of Document mode for applications where users want to keep specific fields and sets of text on specific records within a document. Perhaps many of your customers have such applications. However, those of us who want a long seamless record like Dayflo or Notebook provides aren't well supported by Document mode. My other criticisms of AskSam touch upon features which I feel could be better implemented or which should be added. So far, these additional limitations seem less crippling in making AskSam my TDBMS of choice. G. REPORTS 1. Reports must be split into separate records for header & body. This is a clumsy implementation strategy, but livable. 2. One must have separate reports for directing output to different devices, (e.g., DISK, PRINTER, SCREEN) with commensurate complexity of debugging/revising several different report bodies when the fundamental report logic changes. If a report with just a header could be run in Execute mode, and then follow that report with any other report (body), then this would be much less of a problem (1 body, a header for each output device). Every other DBMS I have seen allows the output device to be selected as an option when printing a report. While AskSam works this way in query mode, only 1 line specifications can be executed in Query mode. Since most formatted outputs that I need require several line specifications, I have to use Execute mode. 3. I would like to have distinct paragraphs of text formatted in a report. Image mode will respect paragraph boundaries, but will chop off the right portions of a line if I try to print into a narrow column on the right side of a page. Stream mode eats up all the spaces it sees. In the short run, I guess that I might place separate paragraphs in separate records - a kludge. I think that Stream mode should recognize a paragraph marker - like 2 carriage successive returns. 4. Like Dayflo, AskSam supports multiple values for a field. However, AskSam seems to insist that multiple values of a field will be printed in Column 1, regardless of the margin settings. I'm not sure if this is a bug or design flaw. The user should be able to control the placement of multiple values of a field and the margin settings should be the guide to be consistent with other AskSam reporting features. 5. I must throw a blank page with Top to get page numbering straight. This means that a 2 page report requires four pieces of paper on my Epson printer. 1 page for AskSam to get its count straight, 2 pages for the report, and 1 page for the Epson printer to allow a page to be ejected. I can't expect Epson to redesign their hardware so as to retrofit printers. But AsKSam need not start pages at 0 before an obligatory blank page is thrown. The Number command should take an argument like Number 1 to start numbering on page 1, Number 10 to start numbering on page 10, etc. 6. It should be possible to place page numbers and dates on the bottoms of pages (Footers), of any Report. 7. Reports don't seem to have an "ending" in which one can simply reset modes at end of report (Sam mode, etc.). Some modes, like margins, stream, can be set and reset in each "report template" this is not a satisfying approach. H. QUERY MODE/UPDATE MODE 1. MANAGING BROWSING I think that AskSam has a good query language but weak browse management. Vincent Puglia's PC Magazine review just does not appreciate the value of views in Notebook. You would improve AskSam if you incorporated some of the features that Notebook's views support. A view is a pointer to a set of records which one selects though some query criterion (or adds to deletes from while browsing). A Notebook view and the Dayflo equivalent help me: a) Know how many records met a search criterion. That's useful to know if I want to narrow my search to expedite browsing and to print a shorter list. b) A view provides a way to jump to the first, last & nth records fetched by a query. AskSam is unduly primitive in NOT reporting the number of records that satisfy a search and by not allowing one to jump to the first or last records with one or two keystrokes. In my books and articles databases, I sometimes want to know approximately how many items match a search criterion. I should not have to count as I browse. If I've retrieved 40 or 50 books/articles and have bounced back and forth in the list of matched records, I'd like to know where I am. (Why should the user of a TDBMS hailed as "one of the most powerful" have to hand simulate a counter when browsing through records?) You can implement these counters and jumps without views, and should. c) Notebook's views also provides a way to keep a pointer to a set of records when one moves across modules (Query, Update, Execute) and also saves having to repeat queries. For simple queries and small data files, the view is less critical. But for as the databases get larger (thousands of records) and searches take longer, a view can save a lot of time. Notebook allows an effectively arbitrary number of views to be saved and accessible at any time.) The use of views reflects the traditional tradeoffs between computational speed and memory use (e.g., always recompute a result or save it for recall). I. Editor: 1. The editor does not allow blocks of text to be moved or copied into multi record documents so that text will have to spill from one record to the next. (See my long discussion above.) 3. I would like a way to make end of field "]" characters sticky so that they can't be accidentally deleted with one keystroke. The ] is a distinguished character in AskSam. I have inadvertently clobbered a ] a few times when altering the contents of a record and then found surprising results in my reports. Having to take care to protect ] characters adds yet another burden for the user. If a ] too 2 presses of a delete key to erase, it would probably save a lot of hassle. 4. The user should not have to commit to move or copy when marking a block. One should copy to a buffer and then move or copy as needed when at the target. J. HELP and the Logic of Seamless Work with Computing AskSam's Help facility is among the worst designs that I have seen. In AskSam, Help can only be called from the Main menu - not from a mode like Query, Update or Execute. The Help file - an AskSam file - replaces the user's file in use. From one angle, this is very elegant. The user can exploit Sam's query language to search the file. One can even customize the Help file with in the Update module. The implementation must have been relatively easy since the Help file is just another file. From my perspective as a user, this Help organization is very disruptive. Suppose that I have been updating a set of 30 records that have matched a query, am on record 25 and want to remember which key will insert the system date into the record (a nice feature). 1. I must leave the Update mode. I will lose a pointer to that 25th record unless I tag it. Since there are no saved views, there is no way for me to save a pointer to the list of records 25, 26, 27, ... that matched the query. 2. I must load the HELP file and save my file. 3. Then I can enter Query mode, type "date" and locate some information that Alt-D will insert the system date into my record at the cursor. 4. Then I must exit the Query mode. 5. Load my original file. 6. Enter update mode. 7. Retype my query. 8. Press the space bar 25 times to find my record x x x x x x x x x x x x x x x x x x x x x x x x x 9. Move the cursor to my previous location within the record 10. Press Alt-D. This is a superb example of DIS-integrated sub-system design. The process of seeking help disrupts the flow of thought and the flow of work. It is probably easier to just pick up the manuals and search the index. In contrast with AskSam, Notebook provides descriptive information with the press of the F1 key within any module. Dayflo provides a standard set of help information that is not context sensitive on the F1-key. This is not great, but at least Dayflo doesn't swap out your file and your location within it. (Your cursor is left its original location after returning from Help.) Dayflo has superbly integrated help information and prompts into the menuing system. If a user presses F1 for help in the course of executing a command, Dayflo provides a Help screen or prompt for that menu option and every subsequent menu option in that command. That "help" supports the workflow. It doesn't distract from it. Every Help facility that I have seen in applications programs -- even clumsy ones -- return the user's cursor to its original location in his original file after Help is completed. AskSam's "Help" is unusual in replacing the user's working file with another file. If AskSam's Help file were simply called an "on-line document" and provided as another file among the other tutorial files on disk I would be less concerned. This help strategy is a fundamentally bad design choice that a person who is trying to design "seamless software" should not pass off to his users. From a marketing angle, you can claim that AskSam provides on-line help. But users aren't really helped by the implementation - they are impeded. (This is another problem at Puglia should have caught, and his oversight makes me distrust the basis of his review even more.) I am unsure how much this help strategy reflects the triumph of marketing over engineering, and how much it is simply a byproduct of the blind spots that develop from designing a mode-based system. In either case the result is similar - a problematic feature that should have not gone past an alpha test. It does not give me faith that subsequent releases of AskSam will move along a developmental trajectory which favors graceful engineering and ease of use. I fear that it suggests that AskSam's future will be cluttered with kludgey "features" that promise assistance and distract in the process. ================================================ In summary, I have found AskSam to be powerful and promising in many ways. But I had overlay high expectations based on the unreservedly rave reviews that PC Magazine gave AskSam. AskSam deserves some praise, but not unrelenting praise. [There are some areas in which designers of other text databases, such as Dayflo and Notebook could learn from AskSam.] AskSam seems compromised by a somewhat disintegrated set of features and limited abilities to handle arbitrary sets of text flexibly. I am most limited by AskSam's clumsiness in text fields longer than 20 lines. But I'm disturbed that a feature as well understood as "help" could be implemented so badly. ------------------------------ From: Yuval Rakavy <yuval%humus.huji.ac.il@RELAY.CS.NET> Date: Mon, 20 Apr 87 00:44:20 jdt Subject: DIAL.SRC Posting of DIAL Program Good day! I am posting here a program that is used to dial a number through using a hayes modem. This program is capable of finding the phone number in a data base. I use this program a lot and I think that many other can enjoy it. Please acknowledge if you get this file as I am not sure that I am using the correct procedure for mailing it to you. Have a good day! Yuval [DIAL.SRC a Unix shell archive is in the lending library. -wab] ------------------------------ Date: 19 Apr 87 8:26 -0800 From: "Ya`akov N. Miles" <bd%dac.triumf.cdn%ubc.csnet@RELAY.CS.NET> Subject: Another Great Product Screwed up by Programming Methodology ATI (a Canadian company) manufactures a short-slot XT-compatible graphics card called the "EGA WONDER" card. It will simulate a CGA, an EGA, or a MONO card to your PC and the mode can be changed in software or permanently in hardware. It will also display EGA resolution on a CGA or a MONO display (with some flicker), and uses VLSI "chips" by NEC. Please be warned that this card comes with a (terminate-and-stay-resident) configuration program called "SMS", which obviates compatibility with the IBM standard EGA card. YA'AKOV Nachum Miles Computer ENGINEER (not scientist) TRIUMF/4004 Wesbrook Mall/ Vancouver/Canada/V6T 2A3 Reply: <bd@triumfcl.bitnet> ------------------------------ Date: 20 Apr 87 06:39:52 PDT (Monday) Subject: Logitech Mouse and Windows From: "DCARIS_Norder.HENR801G"@Xerox.COM First of all, Logitech has, in general, been having problems with some of their driver software. For instance, somewhere around revision 3.10 of their mouse.com, (and related revisions of click.exe and menu.com) they added the capability to "find" menu files via the DOS pathname structure. Unfortunately, they could not handle larger than normal environments, the result being that the system would "hang". Also, Logitech has had trouble working properly with SW that "flips" among different graphics pages. For instance, earlier drivers would not work properly with Codeview. Also, earlier drivers may have had problems with systems running EGA. (Example: Logitech states that release 3.10 now "supports EGA 64K".) Regarding using Logitech mice with Windows: their software includes a special driver for used with Windows (lmouse.drv) which they indicate to "include this in the windows install procedure". If you continue to have problems, I suggest you contact Logitech tech support. (Voice: 415/365-9852 or BBS: 415/364-7057). They have been very helpful in solving problems I have had. Also, the driver software you are using is probably not the most current - even if you bought your mouse directly from Logitech. (For instance, the latest version of mouse.com that I am using, rev 3.12 dated 12/17/86, was provided to me by Logitech tech support.) Regards -- Paul ------------------------------ Date: Mon, 20 Apr 87 13:56:14 est From: tr@flash.bellcore.com (tom reingold) To: brankley%usfvax2@relay.cs.net Subject: Ill Behaved Programs can Trash File System Bob: I had a similar problem last week. It happened to me only once and I don't hope to duplicate it because it gave me such a scare. The symptom was that one of my files was listed thrice when I typed "dir". When I typed "chkdsk", it named my directories and said that they were invalid, offering to change them to plain files! I interrupted the chkdsk and booted. The problem went away. The possible cause of the problem may have been similar to yours. I was programming in C and had tons of stupid errors, like subscripts out of range. Pascal programs will quit when this is detected but this is not the case with C. Instead the array member is accessed, even though it was not allocated. This means that I wrote into memory that did not belong to me. This sort of thing can happen when using dynamic allocation in Pascal, as you were doing. I ran my program several times and that seems to be the thing that caused the damage. Anyway, this is not detailed, but I thought you would find comfort in the fact that I had what may be the same experience. Tom Reingold INTERNET: tr@bellcore.com UUCP: ..!decvax!ucbvax!ulysses!bellcore!tr ihnp4!mhuxt/ ------------------------------ Date: Mon, 20 Apr 87 11:07:59 PST From: ihnp4!iguana!polyob!Upolyof@ucbvax.Berkeley.EDU Subject: SPL Shareware Announced FOR IMMEDIATE RELEASE 4/6/87 CONTACT Dennis Baer 516 694 5872 Software author Dennis Baer has released the Structured Programming Language, (SPL) a free format block structured programming language that runs on MSDOS and PCDOS operating systems. SPL is an alternative to PASCAL and C. The SPL language is implemented by a translator which converts SPL source code to a Microsoft BASIC program which can then be compiled with Microsoft's Quick Basic, MS BASIC, IBM's BASICA, or ported to machines such as AMIGA, MACINTOSH, ATARI ST, or CP/M and compiled with the BASIC compiler for those machines. SPL has been released as SHARE WARE and is available as file SPLLIB.ARC on BIX in the IBM SIG or available on various bbs systems: 516 334 8221 (1200 baud) No registration required for DOWNLOADING. SPL is also available as volume 666 from PCSIG 800 245 6717. If you have any questions you may call me at 516-694-5872 from Monday thru Friday from 10:00 am to 6:30 pm New York time. Some major features and advantages of SPL o SPL is an alternative to the PASCAL and C languages o SPL programs can be run on MACINTOSH,AMIGA,ATARI ST,CP/M o The SPL processor will run on MSDOS emulators on MACINTOSH, AMIGA,ATARI ST o PROCEDURES o WHILE loops o FOR loops with REAL and INTEGER indices and increments o REPEAT loops o Powerful IF THEN ELSE constructs o Powerful RANDOM and SEQUENTIAL INPUT/OUTPUT including formatted OUTPUT o GRAPHICS statements PSET DRAW LINE CIRCLE PRESET SCREEN ..... o BEGIN END blocks o ERROR trapping o Statement labels (multiple labels supported) o Strong data types INTEGER REAL STRING scalars and arrays o Names of variables and labels up to 40 characters upper and lower case o Supports mathematical functions SIN COS TAN LOG EXP ..... o STRING functions MID$ LEFT$ RIGHT$ STR$ VAL$ ASC$ ..... o Your compiled BASIC programs do not become obsolete link them together o SPL programs run faster than PASCAL programs o SPL programs can take advantage of an entire 640k IBM PC o The SPL processor will work on an IBM PCjr with 128k and 1 drive [This looks like it got truncated. -wab] ------------------------------ Date: Wed, 15 Apr 87 09:46:05 pst From: mipos3!cpocd2!rod@Sun.COM (Rod Rebello) Subject: Does TDEBUG support Turbo Pascal V3.02a? >TDEBUG.ARC is available on our CLub Bulletin Board: SE CT IBM PC USER'S >GROUP RBBS. (203) 886-5265. 300/1200/2400 baud, 24 hrs, 7 days. No fees. >Log on and get registered. Start down loading right away. Do either of these versions support Turbo Pascal V3.02a? I have gotten a copy from a local BBS which only supports up to 3.01a. I am getting desperate for a good debugger. Rod Rebello ...!intelca!mipos3!cpocd2!rod ------------------------------ Date: Fri, 17 Apr 87 11:29:15 CST From: BFDI516%UTA3081.BITNET@wiscvm.wisc.edu Subject: Accelerating the Leading Edge Hi I have a Leading Edge, I bought it just before they came out with the turbo versions. I was wondering if anyone knows about the options I have to increase it s speed? I was wondering about the v20 and v30 chips and the accelerator boards compatibility speed, price and effectiveness? thanks harold bitnet address bfdi516 at uta3081 ------------------------------ Date: 17 Apr 87 21:08:00 EST From: "NRL::WILSON" <wilson%nrl.decnet@nrl.arpa> Subject: Honeywell VIP7300 and/or 7800 Terminal Emulators Does anyone have experience using communications packages that will emulate the Honeywell VIP7300 or VIP7800 terminals? We are looking for the names of software and comments on ease of use. THANKS. Alwilter T. Wilson Naval Research Lab 4555 Overlook Ave. SW Washington, DC 20375 (202)767-1343 ------------------------------ From: jsweet%icse.UCI.EDU@ICSE.UCI.EDU Subject: DOS 2.1 and Disk Space Date: Sat, 18 Apr 87 00:29:02 -0700 I want to install a 30MB disk in a true-Blue PC/XT. It runs PC-DOS 2.1 and MUST continue to run that version. Don't ask me why. Now, I have heard two scare stories about the way PC-DOS 2.1 deals with lots of disk space: Scare story one: PC-DOS 2.1 blows up as soon as 16MB of a disk has been used. Scare story two: PC-DOS 2.1 blows up as soon as the space consumed by a single data file exceeds 16MB. Which, if either, of these scare stores is correct? -jns [Don't know about those stories but DOS 2.1 does use large disks less efficiently. You might do better to run 2 16MB segments for this reason alone. -wab] ------------------------------ Date: Sat, 18 Apr 1987 09:00:10 MST From: Villy G Madsen <VMADSEN%UALTAVM.BITNET@wiscvm.wisc.edu> Subject: Display Write 4 with Laser Jet Plus We are having all sorts of grief trying to interface DW4 to an HP Laser Jet plus. Has anyone else experienced problems and then resolved them? A direct answer, rather than through the digest would be appreciated, just so I get it sooner. Thanks ------------------------------ Date: Sat, 18 Apr 87 09:01:53 pst From: reynolds@ames-prandtl.ARPA (Don Reynolds) Subject: Control-Shift-Alt-key Combinations Was surprised to find my new XT clone (with ERSO BIOS) would not insert the ASCII codes 128-255 into PC-Write version 2.71 using ALT plus the keypad. It did work for such primitives as COPY CON [filename] and EDLIN, but I guess I'm spoiled. It may have worked under PC-DOS 3.2, but would not under 2.0 and 3.1, and I like the "dater" and "timer" patches from PC Magazine in PC-DOS 3.1. (Anyone have these for DOS 3.2 and/or 3.3?) Swapping for an AMI BIOS seems to be a win. But for future reference, I'd like to know if anyone has written a program to request all keyboard combinations (Control-Alt-Plus, etc.), and record the scan codes. If programs are written assuming an IBM XT with "standard" IBM keyboard (whichever one you choose to be "standard"), one would have guidance implementing key reassignment with ANSI.SYS, NANSI.SYS, or patching KEY???.COM. Maybe PC-Write is one of the few programs that uses multiple Control-Shift-Alt key combos, but has anyone hit this one before? Thanks. Best, Don (415)694-4027 P.S. This Taiwan clone was supplied with what might be the TIMER.COM program requested in this digest. ------------------------------ Date: Sat, 18 Apr 87 23:36:49 MEZ From: Erich Neuwirth <A4422DAB%AWIUNI11.BITNET@wiscvm.wisc.edu> Subject: 8087-80287 Problem I have a problem with my ATclone and the 80287. I have an 80287 installed and not all programs find it. IBM Advanced diagnostics find it installed. CPUID from CCUC informs me that the 80287 is installed but that I have to flip switch 2 in block 1 in my motherboard. My dealer informs me that there is no such switch block and therefore no such switch. He also tells me that original IBM ATs also do not have such a switch but that all this information is stored in CMOS RAM. Is there a possibility that I can simulate this switch (or do I have one and only am unable to find it)? By the way: CHECK87 also from CCUC informs me I do not have a numeric coprocessor installed. Please answer directly also if possible. I am A4422DAB at AWIUNI11 in BITNET Erich Neuwirth ------------------------------ Date: Sun, 19 Apr 87 23:38:19 ast From: mayerk@eniac.seas.upenn.edu (Kenneth Mayer) Subject: Disk Recovery Programs This is my first contribution to this bboard. I am looking for recommendations on disk recovery software. In particular, I am interested in programs that will work on floppies, and (possible a different program) hard disks. Price is a lower consideration than effectiveness at recovering crashed disks. If this has been asked before, I would appreciate a pointer to the relevant digest article. Thanks, Ken ------------------------------ Date: Mon, 20 Apr 87 15:47:35 MEZ From: Erich Neuwirth <A4422DAB%AWIUNI11.BITNET@wiscvm.wisc.edu> Subject: Graphics toolbox for TurboPascal wanted I would need a graphics toolbox for TurboPascal fulfilling the following requirements: Running with CGA Fill with at least 5 different shades of gray (possibly also in colors) for closed polygonal regions. Borland's Grafix toolbox only has fill for rectangles. Does anybody out there know such a product or does something of that kind even exist in the public domain? Please CC your answer to me directly also. A4422DAB at AWIUNI11 in BITNET. Erich Neuwirth ------------------------------ End of Info-IBMPC Digest ************************ -------