Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!jato!hanauma.jpl.nasa.gov!hyc From: hyc@hanauma.jpl.nasa.gov (Howard Chu) Newsgroups: comp.sys.atari.st.tech Subject: Re: MACH for a TT Message-ID: <1991Mar28.083305.11888@jato.jpl.nasa.gov> Date: 28 Mar 91 08:33:05 GMT References: <1991Mar12.150549.18293@actrix.gen.nz> <1991Mar12.223746.6287@zip.eecs.umich.edu> <4bwC1C200Uh742L317@andrew.cmu.edu> Sender: news@jato.jpl.nasa.gov Distribution: all Organization: SAR Systems Development and Processing, JPL Lines: 47 Nntp-Posting-Host: hanauma.jpl.nasa.gov In article <4bwC1C200Uh742L317@andrew.cmu.edu> mc4c+@andrew.cmu.edu (Mark Choi) writes: > I would like to cast my vote for someone(s) doing this port. Mach is >a NeXT :^) generation operating system, and as such, I would like to see >Atari join Steve jobs and CMU on the cutting edge. If we could only get >NeXTStep.... Bleah. As you may recall, the original NeXT cube used a 25MHz 68030 and MACH. As you may also recall, the original NeXT cube ran at blazing slug-like speeds. Even with 32MHz 68030 in the TT, and assuming 80-90% cache hits, I don't think MACH would be tolerable on the TT. BSD 4.4 would probably be a better choice, it at least is building on a proven base. Steve Jobs is hardly on the cutting edge... As for CMU on the cutting edge - how can you possibly say that with a straight face, with your account on Andrew? Ok, all ribbing aside... Why MACH? OSF has abandoned it. For Unix today, there's only a few main things to consider, and MACH isn't one of them; it doesn't have any major industry support with OSF withdrawing. OSF, whatever that will look like, if it gets done any time soon, SysVR4, which is available now, and 4.4 BSD, which is in the wings. Sun and AT&T are (naturally) really big on SysVR4, current SunOS stuff is trying harder and harder to be SysV instead of BSD based. (And it's annoying as all hell, let me tell you. }-) BSD 4.3, surprise surprise, passes the Sys V validation suite, so it looks like most of the differences (BSD 4.3 vs SysVR4) are mostly academic. BSD 4.4 offers, among other things, native support of the ISO networking protocol stack, which would seem to position it Very Well as the system of the future, as well as the best system for Right Now. Who knows... (Was I trying to sound authoritative there? Ha, fooled you. I just don't believe opinions should be humble... }-) What are the issues? Performance and compatibility. What does "compatibility" mean? Binary compatibility? (Who is there to be compatible with - the Amiga? NeXT? Not the Sun, SunOS isn't enough like any of the standard systems anyway.) Source compatibility? If the various supporters' claims are true, all the contenders are source-code compatible with each other, or more specifically, with BSD 4.3. On another subject... It always makes me laugh when I read about these "real workstations" and their "graphics hardware," with umpteen zillion features and "double-buffered graphics!!!!" Wow! Big deal! A 4 meg ST will let you have *128* buffers, sliced up any way you want! The ol' 8bit let you slice up a screen display on a per-scanline basis. Talk about manipulation power! Geeze, these "Real Computer" makers could learn a thing or two from some "dumb game machines" ... Ah for a hardware smoothscroll on the machines I use at work... -- -- Howard Chu @ Jet Propulsion Laboratory, Pasadena, CA Disclaimer: How would I know, I just got here!