waltrip@capd.jhuapl.edu (02/08/91)
In article <8NuZw5w163w@tz.wimsey.bc.ca>, johankha@tz.wimsey.bc.ca (e_mou) writes: > Well if IBM is going to solely develop OS/2, then we can just send it out of > the window and forget about it. Too bad, it could have been a decent > environment. Windows needs multithreading. Badly. Someone posted a statement to this newsgroup that MicroSoft was no longer developing OS/2 and it has been repeated a number of times since then. I don't believe it's true, although MicroSoft's policy has been so confusing that you couldn't blame anyone (including me;^) for whatever conclusion they drew. MicroSoft appears to be developing a portable OS/2 that, last I heard, was supposed to support Windows, PM and be POSIX compliant and be available in a couple of years. Now we all have our opinions about what that may be worth, but what it implies for the marketplace is clear in at least one important respect: there will be multi-tasking OS to host all of those Windows apps that are being developed out there and the Windows environment will have to be reckoned with for years to come UNLESS there's a big change in the requirements of the user environment (e.g., voice recognition/response). In that case, an operating system that was truly portable, had multi-user protections, could launch threads on multiple processors including special purpose signal processors and included a programming environment where code development using re-usable objects was encouraged just might have a chance to prosper. Now who might have something like THAT by then? c.f.waltrip Internet: <waltrip@capsrv.jhuapl.edu> Opinions expressed are my own.
glang@Autodesk.COM (Gary Lang) (02/14/91)
> whatever conclusion they drew. MicroSoft appears to be developing a > portable OS/2 that, last I heard, was supposed to support Windows, > PM and be POSIX compliant and be available in a couple of years. isn't it becoming obvious that OS/2, when it finally becomes a mainstream OS will be a Unix variant? Why else would Bill buy 20% of SCO? What else makes sense, spend another 20 years debugging a new operating system? Their last go-round on this took them 4 years and the result was an operating system that had Windows with a different imaging system that noone is supporting but IBM on top of a reasonable but new (i.e. un-debugged in the field) operating system that didn't even have a new file system. You heard it here first: Both MS and Apple will be selling a Unix variant under the hood of Windows and the Mac within 3 years and we can get on with the show. In the meantime, NeXT has already seen the light and built a great computer in this mold. This has been my opinion for 6 months or so, and the latest moves to merge the efforts of MIPS, MS, SGI, and Compaq only serve to strengthen my belief in this scenario. One thing I wish IBM would do is realize the marriage is off and take their commitment to NS to the uh, next step and purge PM from their strategy and simply use NS as the interface of choice on all of their platforms. This makes far more sense than taking a hunk of 286 assembler away from Microsoft and trying to implement SAA with it. Come on big blue, you know that NeXTStep would run quite well on PS/2's, RS/6000's, and probably even 370's if you'd get off of your duffs and go for it. 1981 is a dim memory, and if you believe that SAA is a Good Idea (and it basically is) then use an operating system (AIX) and an environment provided for somebody a little hungrier than MS so that you can get back on your feet. This would also support your standardization on PostScript for printing. Forget bundling ATM with PM, go for the real thing. Publishers would love you for it. Otherwise, by the end of the 90's, I can easily see MS being a much larger company than IBM. I'm not saying this is bad folks, but if I were IBM, I'd certainly see this as a bad thing. Executive summary for IBM: 1) Take OS/2 and give it a decent burial. 2) Make SAA a reality, and standardize on UNIX (AIX). 3) use NeXTStep as the standard graphic interface, and come up with a reasonable facsimile for character oriented terminals 4) Buy Borland, and take MS on in the applications market. Otherwise, Watson will be spinning in his grave in 10 years. THINK about it. Maybe IBM should acquire little signs that say THINK and plant then all over the company. All IMHO, etc. -- Gary T. Lang (415)332-2344 x2702 Autodesk, Inc. Sausalito, CA. MCI: 370-0730