Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!usc!julius.cs.uiuc.edu!apple!agate!shelby!msi.umn.edu!noc.MR.NET!gacvx2.gac.edu!gacvx2.gac.edu!scott From: scott@erick.gac.edu (Scott Hess) Newsgroups: comp.sys.next Subject: Re: Death of OS/2? (was Re: Microsoft OS/2?) Message-ID: Date: 11 Feb 91 00:30:18 GMT References: <2467@beguine.UUCP> <29814@usc> <70471@microsoft.UUCP> Organization: Gustavus Adolphus College Lines: 105 Nntp-Posting-Host: 138.236.10.8 In-reply-to: edwardj@microsoft.UUCP's message of 5 Feb 91 02:43:45 GMTLines: 105 In article <70471@microsoft.UUCP> edwardj@microsoft.UUCP (Edward JUNG) writes: Don't be surprised to find our future OS to be "object oriented", micro-kernel, and low-overhead multiprocessing; in short, everything that Mach promised to be but isn't (yet) unless you rewrite it yourself. But more importantly, the OS will support applications and end-user needs (like information retrieval and categorization), which is, after all, what an OS is for. If you want POSIX compliance, PM or Windows compatibility, or security, then you'll want OS/2. A long time ago, people postulated that Apple should deliver a secure pre-emptive multiprocessing portable object-oriented protected client-server distributed OS with a "Mac compatibility" layer (eg virtual machine). This is how we are positioning OS/2. [I certainly don't want to get in a flame war - Edward almost always posts decent posts. Just had to comment, though -scott] I think that Apple would have liked to have out a "secure pre-emptive object-oriented protected" [I'm not sure about other parts] already. They are working on it. I believe it's called System 7.0. I believe they've been working on it for, oh, 2 or 3 years now (though it's surely been in progress since before the announcement). Mach isn't too bad. And it is moving, though slowly. It's fairly portable (well, it runs on at least two architectures, Intel i386 and Motorola '030/'040). Most importantly, it's fairly free, and there will be a Unix on top of it eventually. On the one hand, I just want to point out that before damning Mach in favor of Microsoft, we might want to wait until Microsoft has a product which can damn Mach itself, eh? (_How_ long has OS/2 been going to blow away DOS? :-) On another post: gordonl@microsoft.UUCP (Gordon LETWIN) writes: Also, Ajay, who is the single biggest provider of Unix software? A silly little toy-software company called "Microsoft". I would like to point out that this is Xenix, and while Xenix is like Unix, Xenix is not Unix. called "UNIX". I don't pay much attention to UNIX these days, but I seem to recall a few years ago that "the best minds" at AT&T bought some UNIX technology from an outside vendor - a company by the name of "Microsoft". Though AT&T and Microsoft are working together to combine Xenix and Unix. of duffers who "got lucky". Most of our products - even those beneath the dignity of a guru - require a lot of skill and effort. The quality of our talent is at least as good - and often better - than any other place you might be able to name. Come interview us and find out. All products require skill and effort. Else, they seldom are called products . . . And by the way, if UNIX is the result of the smartest OS minds in the field, how come it bites the big one? It has no standard GUI, and it's disk performance is at least 5 times worse then even pathetic old DOS... And HPFS runs rings around the BSD improved file system. DOS is pathetic - that's why it's so efficient. But you might note that DOS doesn't have multitasking, it doesn't have a standard GUI, etc. Of course, the GUI is seldom part of the OS (the Mac is an exception, here). DOS isn't exactly my idea of a fun operating system to work under, regardless of it's speed and all. And the speed advantage becomes less and less as it constrains bigger and bigger processors to run in smaller and smaller memories (relatively, of course). From all the tech stuff I've read about HPFS, it looks to be 80% the same as BSD FFS. It has inodes, it has superblocks, it has amazing allocation to make the interleave nice, indirect blocks, the works. The main problems FFS has is that it has to be flexible enough to handle multiple architectures (CPU, memory, interrupt handling, you get the picture) while HPFS has to handle only one well- defined architecture. An FFS which was specifically tuned to a particular architecture would more than likely have the same performance within a couple percentage points either way. Of course, most companies don't have the resources to do this tuning - Microsoft does. Comes tell me more when OS/2 runs HPFS on Motorola's, eh? p.s.: re bill gates: - a few years ago, as a publicity stunt, Bill Gates challenged a buncha other big name industry types - all programmers - to a public programming contest. He won. Perhaps you should meet the man, Ajay, before you sneer so publicly. Hmm. I think it'd be sort of fun to put Gates in the arena with Richard Stallman. Though I'm sure Stallman would refuse, we can still imagine how it would go . . . But, before saying that Gates is awesome because he beat some "big name industry types", please let us know the names? I personally think Gates is alright - I certainly don't agree with all of his decisions, but that's besides the point. But it would be nice to know who he beat, what were the rules, what they had to write, etc. Saying he "won" is sort of like quoting Dhrystones without giving the source code to the compiler . . . Later, -- scott hess scott@gac.edu Independent NeXT Developer GAC Undergrad "Tried anarchy, once. Found it had too many constraints . . ." "Buy `Sweat 'n wit '2 Live Crew'`, a new weight loss program by Richard Simmons . . ."