Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!sdd.hp.com!news.cs.indiana.edu!ux1.cso.uiuc.edu!cs.uiuc.edu!coolidge From: coolidge@cs.uiuc.edu (John Coolidge) Newsgroups: comp.unix.aux Subject: Re: A/UX concerns (was "A/UX cc -- a ghost from the past") Message-ID: <1991Feb15.071259.21726@m.cs.uiuc.edu> Date: 15 Feb 91 07:12:59 GMT References: <1991Feb13.122652.14565@helios.physics.utoronto.ca> <49119@apple.Apple.COM> <1991Feb14.204018.4351@helios.physics.utoronto.ca> Sender: news@m.cs.uiuc.edu (News Database (admin-Mike Schwager)) Reply-To: coolidge@cs.uiuc.edu Organization: University of Illinois, Dept. of Comp. Sci., Urbana, IL Lines: 95 sysmark@physics.utoronto.ca (Mark Bartelt) writes: >> Well, most of the basic Unix workstation system C compilers are 'prehistoric >> mosters' originating from the terrifying, Tyrannosaurus pcc from >> the continent of AT&T. >Um, I beg to differ. Of the three UNIXes I use most often, two (IRIX from >SGI; Ultrix from DEC) provide ANSI compliant C compilers. I can't vouch >for the fact that they're 100% in conformance with X3J11, but they do at >least support function prototypes. IRIX has had this *at least* since >version 3.2, which is copyrighted 1988; I have no idea how long Ultrix >has had it. I've been using about 7 major and semi-major unix variants for the last few years; most are still using pcc-derived compilers as their base. I agree that the time to ANSI-fy is now; I'm very tempted to say that people shouldn't worry about compilers anymore and just use GCC (or produce an even better freely available compiler :-)); the time for single-vendor compilers is mostly long gone (there are exceptions, such as the excellent MIPS compiler DEC provides with Ultrix, but I suspect that eventually even they will be passed by a compiler that literally hundreds of talented people are banging on). >My real concern is more general than just a complaint about A/UX's cc. >I get the feeling that Apple doesn't really have much of a commitment >to UNIX. Ten or so years ago, DEC had a small (tiny!), but talented >and dedicated UNIX group, before the company even had an "official" >UNIX product. And they had a couple of managers with the foresight to >realize that the company should push forward with UNIX, and the sooner >the better. Unfortunately, most of DEC's management spent most of their >energy trying to fight and suppress the UNIX groundswell (coming both >from customers and from within their own company), rather than making >a strong and unequivocal commitment to supporting UNIX. My opinion is that this is what's been going on in Apple (of course, I'm on the outside looking in, and the people at Apple may see things differently). From what I've seen, A/UX has gone through three distinct stages: 1.0: "The Feds want Unix, and we'll give it to 'em - grudgingly!" I used 1.0 for a very short while; it was quite painful to use for lots of things, and fortunately 1.1 came along quickly. 1.1: "Well, if we have to do it let's at least do it right." A/UX 1.1 (probably should have been 2.0 - it was that much better than 1.0) was a very usable Unix platform. It ran rings around systems from Unix-only vendors and was easily the best OS I've ever seen with a 1.x version number. 2.0: "OK, if we're going to do it right let's go all the way". 2.0 added scads of features to 1.1; the most useful, I think, have been the MacOS mode (I tend to use lots of MacOS tools as well as Unix), shared libraries, and the UFS file system. Unfortunately, as with any feature-driven release, there have been problems. Hopefully 2.0.1 will fix them. Basically what this boils down to is that the A/UX project started off as, basically, a quick hack to comply with federal purchasing requirements. After a while, it turned out that people were actually buying it and using it, and things had to be made usable, so money and people got added to the project and things evolved a bit. Then, after a bit more settling, someone decided that A/UX was a good idea after all and 2.0 was born. If you'd been in the Bay Area last summer, you would have seen full-page advertisements from Apple looking to fill gobs of positions in the A/UX group. Based on that evidence, and the amount of activity and support I've seen on the net, it looks like Apple has decided to make a major, long-term commitment to A/UX. But my guess would be that that commitment started somewhere in early 1989; that's really not a long time to make major steps in upgrading a product. Even as things go, though, our findings have pretty much been that A/UX 2.0 is the most interoperable Unix platform (outside of SunOS, which pretty much for better or for worse sets the standard) we've seen. It works well in a network environment, is very usable and (mostly) stable, and provides a mostly good set of tools. Of course, I also don't touch cc anymore, pretty much ignore as, use the Gnu fileutils since I can't stand some of the SYSVisms, and so forth. But that's also not much of a slam --- I've found at least one or two commands and "features" in other Unixes that I've replaced as soon as possible. All in all, I suppose the biggest complaint we have about A/UX is that it only runs on Apple platforms (:-)) --- there are a lot of faster boxes around here that no one will use unless forced to (and we're not talking MacOS fanatics here, either --- about half of them are Unix-only types who don't like Macs). A/UX is better than most of the Unixes I've seen out there; with some time and effort, I believe it will be one of the best, maybe even _the_ best. --John -------------------------------------------------------------------------- John L. Coolidge Internet:coolidge@cs.uiuc.edu UUCP:uiucdcs!coolidge Of course I don't speak for the U of I (or anyone else except myself) Copyright 1991 John L. Coolidge. Copying allowed if (and only if) attributed. You may redistribute this article if and only if your recipients may as well.