Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!ncar!umigw!umbio!medsoft!p4.f10.n135.z1.fidonet.org!Ed.Maurer From: Ed.Maurer@p4.f10.n135.z1.fidonet.org (Ed Maurer) Newsgroups: comp.sys.ibm.pc Subject: Re: Microsoft, OS/2, and UNIX Message-ID: <268.24375DE1@medsoft.uucp> Date: 2 Apr 89 18:36:00 GMT Sender: ufgate@medsoft.uucp (newsout1.18) Organization: FidoNet node 1:135/10.4 - AMS Support-Net 13, Miami FL Lines: 114 >>Every year the same argument, and every year it gets shot down for a >>few >>damn good reasons: First and foremost, the user base simply will not >>accept >>UNIX in any of its current releases (numerous, and therein lies >>reason 2: > Users will accept whatever is necessary to run their applications. I'd love to have your users -are they for sale? I would guess you've never worked in end-user support. The issue is how much of a fight will they put up? What's the cost in productivity during the learning phase? The PC arena is a far cry from system 3X, where in the old days (yesterday?) there was simply no alternative to mini-applications. Given the choice, many users would prefer a c-64 (it do word processing, don't it?). Theses users must be supported in the same environment as the multi-tasking hot-shot who has used computers since he was a lad in prep school. For most novice and advanced users, however, 'ease' (perceived as much as real) will win over power hands down. >>OS/2 was never designed to be a multi-user OS, and hence any >>comparison to UNIX is absurd to start with. Without a shell, UNIX is >>far >>too much for the average end user - and whose shell is going to be >>the ONE? > I think you mean something like a friendly interface. There are > several standard shells under UNIX which are pretty close the the same > environment for the uneducated user, and which behave remarkably like > the MS-DOS command interpreter until you get to functions which MS-DOS > cannot perform, such as multitasking. Most users can handle logging > in and typing an application's name, much as they can handle turning > the machine on and typing an application's name. What's the problem? Pretty close just doesn't cut it when you're also trying to support the advanced users in the same environment. The MS-DOS command interpreter is a dead issue. The issue is the Presentation Manager and SAA. To date, UNIX simply doesn't fit within the grander scheme of SAA - which is partly the rationale behind the delay in IBM official support, insofar as the corporate response to the concept of SAA has been overwhelmingly positive. (their DP managers, responding to senior management, are still hung up on interconnectivity, remember.) >>Most UNIX based application sites sites.. still require ..[an administator] > You will find that this is changing very fast. How many application > sites must now have a PC expert? Whenever you go to networked PC's > you *need* such a person. Few standalone UNIX boxes require much more > than the initial setup and some simple instructions to run True. But the PC expert is often the secretary who'se been delegated the task. This works with DOS, OS/2 and even Novell, to a point far greater than current UNIX based systems. And look at the exposion in LAN'ed DOS, often implemented despite existing UNIX systems, cost and software availabilty/choice part of the criteria. > You sure must not be trying to run 10Mb in 8Mb of memory. Under UNIX > it's painless. Try it under OS/2 and watch what happens. No, and neither does the average user. Where can you find 10Mb of DOS applications you'd want to run concurrently? > Some people would say that an operating system supported by multiple > vendors is one which is likely to be around for awhile. I heard that argument used as a reason for the imminent demise of APPLE. Most management is well aware that APPLE's single, consistant interface has proven itself in terms of productivity, despite the lack of software. > IBM has announced that AIX for the high-end PS/2 lines is right on > schedule, although it's slipped for some of the higher-end machines, > and that they intend to support it on a very wide range of machines. Although this may have changed, there is no official release date. IBM states very clearly that unless they have specified a release date, they should not be held to any 'intention'. Release dates for all versions of OS/2 have been specified and (wonder of wonders) met. > You have me there. IBM is supporting it. AT&T has been supporting it > supporting it. And we won't even get into all the smaller companies > that are producing versions. What do you consider official? Taken out of context. My statement was specifically meant for IBM. 'Official' is when IBM wants the same standards of AIX support from it's dealers that they expect in DOS/OS2, minimal as they may be. Remember, that since most users, including Fortune 500, are equipped by retail vendors such as Computerland, Entre or Inacomp, it's there that the trench war will be fought. When I can find UNIX or AIX on the shelf at these, or when I spot the first 3-piece on the subway with his 40Mb portable running UNIX I'll buy the argument that there's a real reason to consider it for applications. As of now, you'll find it's OS/2 - not UNIX on the shelves and strap hanging. > OS/2 isn't dead, but neither is COBOL. Take 'em if you want 'em and > let the rest of us get some work done. If by work, you mean CAD, Statistics, or Programming, sure. But for the stand-alone spreadsheets, word processing and database that make the PC world turn, who needs UNIX? If anything, the market is going in the opposite direction - witness the growth of Apple in the fortune 500. Yes, unfortunatly, Cobol is alive and well- and still producing (If it works, who cares?). The question isn't one of UNIX v OS/2 (I think we'll see both for a long time) but rather is UNIX the only/best way to go - so far, I see nothing to indicate validity today any more than 2 years ago, when the same arguments were presented by both sides regarding the proposed OS/2 standards. -- Ed Maurer - via FidoNet node 1:135/3 Medical Software Exchange BBS (305) 325-8709 UUCP: ...uunet!gould!umbio!medsoft!10.4!Ed.Maurer ARPA: Ed.Maurer@p4.f10.n135.z1.fidonet.org