Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!csd4.milw.wisc.edu!uakari.primate.wisc.edu!indri!ames!vsi1!altos86!dtynan From: rk@bigbroth.UUCP (rohan kelley) Newsgroups: comp.unix Subject: "The Clouds Begin to Clear" (long) Keywords: A (dos) lawyer's view of future computing - UNIX wins Message-ID: <12381@altos86.Altos.COM> Date: 31 Jul 89 21:45:24 GMT Sender: dtynan@altos86.Altos.COM Organization: UNIleX Systems, Inc., Fort Lauderdale, FL Lines: 222 Approved: dtynan@zorba.Tynan.COM The American Bar Association has as one of its sections, a section on economics of law practice. Lawyers who have an interest in efficient delivery of legal services may belong to that section. It has many thousand of members. The members of the section are then entitled to affiliate with certain divisions. One of the divisions is the computer division. Within that division, there are many sub groups which include user groups such as mini/mainfram user group, document assembly user group, communications UG, word processing UG, litigation support systems UG, etc. There is even a small (but elegant and dedicated) UNIX USERS GROUP. (In fact, I hope that some of its members who are on the net will post followups to this so you can see that unix lawyers do really exist.) The legal market for hardware and software is a substantial market and there are many lawyers who are both interested and in and knowledgable regarding computers. Anyway -- the computer division of the economics section publishes a periodic news letter called "Network". In the May/June issue there appeared an article written by James Sprowl. >From the article, I surmise that Jim is a dos prisoner. However, since he is writing for an audience of (mostly) other legal dos prisoners and since he views the future as seen through the eyes of a non-unix person, it strikes me that his views as published and read by thousands of lawyers who have a strong interest in computers generally might be of interest to netlanders and to the companys who employ them. There are technical inaccuracies in the article, however it should not be criticized for that. Its value lies in seeing the greater horizon from the eyes of a dos prisoner who is looking beyond his prison bars. It is of significant interest to note that he concludes that "OS/2 with the Presentation Manager will emerge first and will prosper for a time, but I predict that UNIX will be the big winner in the long run." Some of you may also have noticed the article in the latest Word Perfect newsletter which extolls the virtues of the UNIX operating system and gives a good historical overview. Of course there is also the infoworld article from two weeks ago which test and compare Xenix and Interactive SysV. If the mainstream dos user recognizes that there is new and exciting software (os) technology out there which can take them far beyond their current limitations, that is significant. That they don't know that UNIX was there years before dos was created (nearly before Bill Gates was born) isn't important. With permission of the author, I have reproduced his article which I believe you will find interesting even if you don't agree with all his conclusions. ======================================================================= Rohan Kelley -- UNIleX Systems, Inc. (Systems and software for lawyers) UUCP: ...{gatech!uflorida,ucf-cs}!novavax!bigbroth!rk (office) novavax!mdlbrotr!rk (home) ATTmail: attmail!bigbroth!rk 3365 Galt Ocean Drive, Ft. Lauderdale, FL 33308 Phone: (305) 563-1504 "Go first class or your heirs will" -somebodyelse ======================================================================= Sprowl Predicts Where Hardware is Headed: The Clouds Begin To Clear By James A. Sprowl, Technical Editor, ABA506 When purchasing a computer today, you have to make a difficult choice. Do you purchase the least expensive computer and risk it not being fast enough or powerful enough to run the programs of tomorrow, as well as some of those of today? Do you purchase a faster mid-price model that is more likely to run all the new programs for the next few years but which may not be able to handle the truly advanced programs now appearing on $10,000 systems or the upgrades of today's programs? Do you purchase a "top-of-the-line" product now, knowing that the software to fully utilize the product's power is not yet available and that the same computer will cost much less in a year or two? Or do you purchase an Apple Macintosh to gain the most advanced user interface at the price of possibly cutting yourself off from some of the software options? These are difficult choices. And until very recently, no one could give you any indication of what the future would hold. Would the IBM operating system of the future be MS-DOS, UNIX, or Os/2? What of the NEXT computer from Steve Jobs, the founder of Apple? Would IBM compatibles still be around in a few years? What about Apples? Prior to this past December, these questions could not be answered. But now information released by the major players in the marketplace has finally given us some indication of where things are headed. Let me piece together the information recently released, try to predict where we will be several years from now and suggest what impact all of this should have on your purchasing decisions in the months to come. (But my wife would warn you that I bought into NBI stock at 31; it is now selling for 1-1/2.) First of all, the Macintosh system is still several years ahead of the rest of the industry. Unless Apple really botches things, the Mac will remain the standard against which other computers are judged for some time to come, and its sales will continue to grow. It will remain a premium product, priced 50 percent above the price of comparable IBM PC compatible equipment, and it will continue to build market share against the competition. But because there are more PCs than Macs in the field, software development will continue to lag on some Mac Products. (Regarding system prices, see "A Systematic Computer Systems Decision," page 1. -- Ed.) It is now clear that OS\2, IBM's and Microsoft's vaunted new operating system, has failed to catch on in its present form, even with its fancy Presentation Manager user interface. Therefore, it will be redesigned extensively over the coming 12 months into a much more powerful system that will work much more like the Macintosh user interface. Look for the following improvements: "point-and-click" mouse facilities, very large file and disk drive memory management capabilities and support for very long file names. The tradeoff for these improvements will be far less compatibility with PC-DOS and MS-DOS. Os/2 files will no longer be compatible with DOS files, although both types of files will coexist on the same disk drive. When using OS\2 on the newest 80386 and 80386sx PCs a year from now, you will be able to run two or more standard PC-DOS programs at the same time. The price? You will need a fast 80386 PC with 4 to 8 megabytes of expensive high-speed memory. That 80286 PC you bought because you were promised it would run OS\2 will run OS\2 -- the old version of OS\2 that Microsoft and IBM will begin to leave behind next year when the new 80386 version of OS\2 begins to take over the universe. You will be able to network your 80286 machines and 80386 machines together using this older version of OS\2, but the future otherwise belongs to the 80386 machines. The problem with the 80286 processor chip is historical. Like the 8088 it cannot manage more than 64K of program or data at a single time. It forces programmers to slice up programs and data into little bitty pieces. It also cannot intercept screen writes and therefore cannot "tame" many DOS programs into running within a window. And the 80286 can only run one DOS program at a time, while halting everything else. The 80386 processor chip, all the Macintosh (actually Motorola 68000 series - Ed.) chips and the new, much faster RISC chips can manage much bigger chunks of memory. They do so by slicing up the chunks in ways that the programmers are not even aware of. In the long-term, the programmer community will "doom" the 80286 processor (this will probably take five years or more, however). Meanwhile, UNIX continues to gather support in the marketplace, with the federal government insisting on UNIX minicomputer versions of all programs it purchases on PCs. AT&T and several other vendors are rapidly working to integrate all versions of UNIX into a single operating system that can be ported over to virtually any central processing unit. And, under funding provided by IBM and DEC, MIT has created a public domain graphical windowing front-end for UNIX called "X Windows." X Windows is supported by IBM, Apple and all the players. It creates a uniform, networkable user interface into which other standardized user interface packages may be readily installed. The beauty of UNIX with X Windows is that the dissimilar hardware of different computers and terminals is hidden behind X Windows -- X Windows looks the same despite the user's hardware. This means that any UNIX user running X Windows can use any one of several different UNIX user interfaces that will be developed over the next year. What will the UNIX user interface be? AT&T, which has now teamed up with Sun Microsystems, the most successful maker of professional workstations, has proposed a UNIX user interface standard: "Open Look." It will be shipped with AT&T Standard UNIX starting this Fall. But Open Look may not catch on. Concerned about the power of AT&T teamed up with Sun, other vendors, lead by IBM, DEC and Hewlett Packard, have set up the "Open Software Foundation" to develop their own standard version of UNIX. The Open Software Foundation recently adopted Microsoft's Presentation Manager as a UNIX standard user interface. Think of it -- the next incarnation of Presentation Manager may become a standard user interface for both OS\2 and UNIX! I can't think of anything that would make software vendors happier than to have the same program source code run substantially without change under both UNIX and OS\2, and also under MS- DOS with Windows 386. So this proposal is a natural to catch on quickly with the DOS and UNIX software community. Another candidate for the standard user interface for UNIX is NEXT computer's NextStep interface, which is based on the PostScript language that has taken over the high-end printer market. Look for NextStep to grow in popularity. Software developers will like it since it permits the same programmer language to be used when talking to the printer or screen, and it will lend itself nicely to desktop publishing. But software developers will initially prefer the Presentation Manager interface. Accordingly, I predict that UNIX with the Presentation Manager will win the operating system wars, long run. It will run PC-DOS programs in windows on 80386 PCs, just as early PCs ran CP/M programs for a few years until CP/M programs went out of style. You will need an 80386 IBM PC or compatible, or an 86030(sic) Macintosh or a new RISC processor (see below) to run this new version of UNIX, which will be ready in about a year. It will catch on slowly. OS\2 with the Presentation Manager will emerge first and will prosper for a time, but I predict that UNIX will be the big winner in the long run. And since software houses will be able to port programs freely back and forth between UNIX and OS\2, there will be no shortage of software on either operating system, and they will share the limelight for some time to come. OS\2 will always require an 80386 or 80486 -- a particular type of microprocessor. Even Intel now admits that the newer RISC chips are much faster than the 80386. Intel just introduced what may be the worlds fastest RISC chip, and it will never run OS\2 -- but it will soon run UNIX. UNIX will also create a bridge for software between the Mac and IBM/Microsoft worlds. The 80286 IBM PC compatibles and the present version of the OS\2 operating system are already obsolete. If you purchase any 80286 machines, you should shop for price and plan on giving them a less prestigious role in future years. Do not shop for speed that is not needed now and that will not help you in the future -- the new programs will not run properly on this microprocessor. Personally, and to be economical, I plan to stick to the most inexpensive of the 10-megacycle conventional 8088 PCs. If I were into networking, however, I would stick to the 12-megacycle 286 PCs which will run OS\2 for networking. I would avoid 386s for the next year or two (except where their speed is needed) until OS\2 for the 80386 emerges from Microsoft's laboratories in about 12 to 18 months. At that time, the price difference between 286s and the 386s will be so small there will be little point in purchasing the 286s. The 386s will be available at that time in all price and speed varieties. And the 386s will readily adapt to UNIX when it arrives. Meanwhile, if you want the best in sophistication, buy a Macintosh with a 68030. It is here now, and it will adapt to the future whenever the future arrives. But make sure it supports all the software that you need. What of displays? Only the VGA standard, or extensions of VGA standard, will survive into the future on IBM PCs and compatibles. What of printers? Longterm, only the PostScript equipped printers will survive. PostScript printers will sell at premium for several more years. James A. Sprowl, 135 S. LaSalle St., Chicago, IL 60604. 312/372-7842. ABA506.