Return to BSD News archive
Newsgroups: comp.unix.bsd.freebsd.misc Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.rmit.EDU.AU!news.unimelb.EDU.AU!munnari.OZ.AU!news.hawaii.edu!ames!lll-winken.llnl.gov!uwm.edu!news.inap.net!news1!not-for-mail From: root@dyson.iquest.net (John S. Dyson) Subject: Re: SUN moving away from BSD to System V X-Nntp-Posting-Host: dyson.iquest.net Message-ID: <4lf01d$ar@dyson.iquest.net> Sender: news@iquest.net (News Admin) Organization: John S. Dyson's Machine References: <4lce4k$flh@news.umbc.edu> Date: Mon, 22 Apr 1996 03:56:29 GMT Lines: 36 In article <4lce4k$flh@news.umbc.edu>, Sandip Srivastava <ssriva1@umbc.edu> wrote: >A friend of mine who does quite a bit of C and Unix programming on Suns, >told me that SUN is moving away from a BSD based system towards System V. > >And isn't Linux based on System V and POSIX compliant? > Linux is NOT based on System V -- that is a common misrepresentation. It is a total rewrite. One distribution of Linux is POSIX certified, but for example AFAIK, Red Hat, etc (the common ones) are NOT. I used to maintain SVR4 for a major computer manufacturer. Except for some files being different in different places, when I USED SVR4, it wasn't that different from FreeBSD (FreeBSD also blew SVR4 away in perf.) I was also responsible for writing QC code and at times running benchmarks, and would test FreeBSD as a baseline. The day-to-day management of the SVR4 box was very different from both FreeBSD and Linux. For example, many good SVR4+'s have Veritas file systems -- I don't think that FreeBSD or Linux has that yet (for better or worse, probably never will.) Also, for neatness of the kernel (I's dotted and T's crossed) out of FreeBSD or Linux, NetBSD probably has the edge :-). For performance under load FreeBSD has the edge (given the same generations of code.) For snappy interactive performance (mostly due to async metadata), but some loading anomolies, Linux is pretty good. I think that the reason that Sun is going to the SV base, is that it is finally getting to be good enough. If SVR4 was originally done well, it would have been SUPER... The original SVR4's looked (internally) like a research project on steroids -- scarier than the stuff in /usr/src/linux in most places, but at least they used manifest constants most of the time instead of "magic numbers..." I haven't seen 4.2 or Solaris 2.5, so my expertise is very limited there. John