AWCTTYPA@UIAMVS.BITNET ("David A. Lyons") (06/21/88)
>Date: Mon, 20 Jun 88 00:31:35 EST >From: Murph Sewall <SEWALL%UCONNVM.BITNET@mitvma.mit.edu> >Subject: Re: DOS/ProDOS >The point is I am NOT anti-ProDOS. Everything I've seen indicates it >is a MUCH more sensible system for developing applications, but a lot >of this "ProDOS is the ONLY way to go" chatter seems to be coming from >people who arrived (and bought their software) after 1985. True, ProDOS isn't the ONLY way to go. If there's a reasonable decision to make, I always say to go with ProDOS. Certainly very little NEW software is being developed for DOS 3.3, and it is gradually dying out. New stuff like AppleShare (file server) does not support DOS 3.3, and as storage devices continue to grow, it gets less and less reasonable to keep using DOS 3.3. CD-ROMs would be ridiculous, for example, addressed 400K at a time. >How come there are so many ProDOS word processors and not one decent >ProDOS text editor (WYSIWYGs make TERRIBLE text editors and besides >they hog core with lots of "features" that a word processor needs but >a text editor doesn't)? Well, there's FreeWriter, and the ProDOS AppleWriter (is it still being sold?). Are those the kind you're looking for? I suspect we will see some more text editors being developed. (I wouldn't mind writing one some day, and some of my friends are inclined in that direction, too. Writing their own, I mean, not just wanting me to write one!) >Though I like hard disks for working storage and software, they >aren't reliable enough to trust anything truly important to (hence >the constant admonishions to back up the rascals). Finished work >belongs on disks (plural) anyway. I can keep track of volume numbers >as easily as pathnames, so two 400K volumes on a 3.5" disk isn't >really a handicap. My RAM-Charger just died; while RAM disks are >truly neat, they are even less reliable than hard disks. That's not very fair--I don't find hard drives to be less reliable than other storage devices. The reason backing up HDs is so important is that you can lose a LOT of stuff all at once. If your power goes out while writing to the main directory on the hard drive, you can lose everything on it (unless you want to spend a lot of time recovering it and PROBABLY succeeding); it's not so big a deal on smaller devices, since there's less data involved. You should ALWAYS back up everything important. I've had trouble with 3.5 drives, 5.25 drives, RAM drives, and hard drives. But I back things up, and I've never lost much work. [Sorry--I've lost track of who ">>" is; the person Murph was quoting] >>2. no support for modern hardware such as eighty column displays, >> extended memory and interrupting devices such as the mouse >Funny how well my 80 column card works with my DOS 3.3 applications; and >I've got a number of 128K DOS 3.3 programs. ProDOS supports the 80-column card to exactly the same degree that DOS 3.3 does: NOT AT ALL. 80-column support is completely up to the applications you're running. >> 4. DOS has no consistent call interface; as a result, DOS >> was only reassembled once in its five-year life, as this >> would cause routine entry points to change, causing >> programs to quit working. As a result, all bug fixes >> consisted of applying patches to the existing code. >That's more esoteric than I care to investigate, but it strikes me as >peculiar that Diversi-DOS, David DOS, One-Key DOS, and ProntoDOS work >so well. I don't believe any of those are merely patches to the >existing code. Actually, all of those ARE patches to DOS 3.3. *NOT* minor ones, but still patches. I wouldn't call them "mere" patches--they are certainly significant and complicated. >Programmers sometimes use undocumented entry points resulting in no >end of trouble when the operating system changes [...], but it >doesn't seem to be all that bearish for a really good systems >programmer to keep the documented entry points where the >documentations says they'll be. Unfortunately, very little of DOS 3.3 was officially documented. Apple documented the RWTS interface, which could read and write SECTORS on disks. It didn't provide any access to the file structure. The File Manager (FAM) was documented in places like Beneath Apple DOS. It is a fairly clumsy interface, and it wasn't really sufficient to do all the things programmers wanted to do. So they quite understandably started using other, undocumented entries to DOS. The ProDOS MLI (Machine Language Interface) is slick, well-documented, and leaves no good reason for programmer to code around it. >ARPA: sewall%uconnvm.bitnet@mitvma.mit.edu Murphy A. Sewall >BITNET: SEWALL@UCONNVM School of Business Admin. >UUCP: ...ihnp4!psuvax1!UCONNVM.BITNET!SEWALL University of Connecticut --David A. Lyons a.k.a. DAL Systems PO Box 287 | North Liberty, IA 52317 BITNET: AWCTTYPA@UIAMVS CompuServe: 72177,3233 GEnie mail: D.LYONS2
SEWALL@UCONNVM.BITNET (Murph Sewall) (06/21/88)
>It gets less and less reasonable to keep using DOS 3.3. CD-ROMs would >be ridiculous, for example, addressed 400K at a time. True, but 32Mbytes at a time may look just as silly on a 1+ gigabyte device (I know ProDOS can be modified to handle it -- GS OS again? -- but I couldn't pass-on the thought 'cause I'm not so sure from what I've read that DOS 3.3 couldn't have been modified as well <too many other problems to make that I worthwhile approach, no doubt>). >>How come there are so many ProDOS word processors and not one decent >>ProDOS text editor (WYSIWYGs make TERRIBLE text editors and besides >>they hog core with lots of "features" that a word processor needs but >>a text editor doesn't)? > >Well, there's FreeWriter, and the ProDOS AppleWriter (is it still >being sold?). Are those the kind you're looking for? Not really; FrEDwriter will do in a pinch, but it's paragraph oriented (fine for letters and manuscripts, but for anything else one must be sure to remember to add the <cr>'s. Mainframes are LINE oriented. What I like about PIE Writer is that it's a line oriented (dot command, but not-WYSIWYG) word processor. So it not only writes files that I can upload without going through the process of printing to disk, but I can DOWNLOAD as well without having to use a utility to strip carriage returns. I wonder who now owns the copyright on PIE's code (it would be awfully nice to get it expanded to 128K and converted to ProDOS). >That's not very fair--I don't find hard drives to be less reliable >than other storage devices. The reason backing up HDs is so >important is that you can lose a LOT of stuff all at once. Wait'll you try and restore from a streaming (mirror) backup tape after replacement of the platter due to a physical crash. No hard disk is perfect, they all are manufactured with some "bad" sectors and the controller is programmed to skip those sectors. Replace the platter and the bad sectors are in a different location. Unfortunately, streaming backups simply reproduce recorded sectors; so if the platter is replaced chances are something important (like part of the FAT) is recorded on what were good but have become bad sectors. The newer, file-by-file, technology solves the problem, but they cost a bunch more too (there's a neat device called "The Fat Boy" described in this week's InfoWorld -- its a 1.2 gigabyte, file-by-file tape backup that uses standard Video 8 cassettes and transfers at 15 Mbytes per minute). Optical, or even optical-magneto, storage which will begin appearing this Fall is potentially more reliable than any of the mass storage devices available at the moment. --------------------- Disclaimer: --- My employer isn't responsible for my mistakes AND vice-versa! (subject to change without notice; void where prohibited) ARPA: sewall%uconnvm.bitnet@mitvma.mit.edu Murphy A. Sewall BITNET: SEWALL@UCONNVM School of Business Admin. UUCP: [rutgers psuvax1 ucbvax & in Europe - mcvax] Univ. of Connecticut !UCONNVM.BITNET!SEWALL "It might help if we threw all the MBA's out of Washington." - Adm Grace Hopper
ralphw@ius3.ius.cs.cmu.edu (Ralph Hyre) (06/25/88)
In article <8806210243.aa00242@SMOKE.BRL.ARPA> AWCTTYPA@UIAMVS.BITNET ("David A. Lyons") writes: >The ProDOS MLI (Machine Language Interface) is slick, >well-documented, and leaves no good reason for programmer to code >around it. Well, except for performance, perhaps - but you could argue that that's Apple's problem. The kinds of thing I'm interested in right now (MIDI at ~38.5 kbps, fast serial communications) would fairly tax the ProDOS interrupt handling mechanism. -- - Ralph W. Hyre, Jr. Internet: ralphw@ius2.cs.cmu.edu Phone:(412)268-{2847,3275} CMU-{BUGS,DARK} Amateur Packet Radio: N3FGW@W2XO, or c/o W3VC, CMU Radio Club, Pittsburgh, PA