Return to BSD News archive
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!spool.mu.edu!howland.reston.ans.net!swrinde!newsfeed.internetmci.com!EU.net!Portugal.EU.net!news.rccn.net!newsfeed.sunet.se!news01.sunet.se!sunic!news99.sunet.se!news.uni-c.dk!news From: pd@kubism.ku.dk (Peter Dalgaard BSA) Newsgroups: comp.os.linux.development.system,comp.unix.bsd.386bsd.misc,comp.unix.bsd.bsdi.misc,comp.unix.bsd.netbsd.misc,comp.unix.bsd.freebsd.misc,comp.os.linux.advocacy Subject: Re: Historic Opportunity facing Free Unix (was Re: The Lai/Baker paper, benchmarks, and the world of free UNIX) Date: 18 Apr 1996 14:21:40 +0200 Organization: Little if any Lines: 54 Sender: pd@bush.kubism.ku.dk Message-ID: <x2ohop3ep7.fsf@bush.kubism.ku.dk> References: <NELSON.96Apr15010553@ns.crynwr.com> <yfgbuktfn1w.fsf@time.cdrom.com> <3172AF35.15A9D29E@wolfenet.com> <31734160.1332595A@lambert.org> <84wx3f2k25.fsf@oberon.di.fc.ul.pt> NNTP-Posting-Host: stat.kubism.ku.dk In-reply-to: roque@oberon.di.fc.ul.pt's message of 17 Apr 1996 11:58:58 +0100 To: roque@oberon.di.fc.ul.pt (Pedro Roque Marques) X-Newsreader: Gnus v5.1 Xref: euryale.cc.adfa.oz.au comp.os.linux.development.system:21811 comp.unix.bsd.386bsd.misc:705 comp.unix.bsd.bsdi.misc:3342 comp.unix.bsd.netbsd.misc:3184 comp.unix.bsd.freebsd.misc:17715 comp.os.linux.advocacy:46344 In article <84wx3f2k25.fsf@oberon.di.fc.ul.pt> roque@oberon.di.fc.ul.pt (Pedro Roque Marques) writes: A common ABI to free Unices. That would be great but... From a user point-of-view what is an operating system but the software that provides the ABI ? Having an agreed upon ABI would make the different kernels nothing but different algorithms of achieving the same OS. The next logical step would be to stop confusing the end user and calling one name to it, no matter what the particular set of kernel algorithms where in use. While in terms of "the desktop war" this would be a clear adavantage, personaly i don't believe the free Unix comunity would accept this. This is not quite true. An ABI is a *subset* of an OS, at least in the common meaning of "OS" (if you equate OS == kernel + libraries, you may have a point, but most of us will also include the set of basic programs supplied). However, there *is* obviously a schisma between "standards compliance" and "product profile" among both freeware and commercial developers. The benefits are that software writers only have to maintain one version and that OS Vendors have an immediate selling point saying that it will run WordPerfect or whatever. The drawback is that every vendor can become alike as you indicate. However that wasn't really what I wanted to say here. I wanted to point out that there actually is a movement towards a common Intel ABI, called ABI+. The new thing in it is that they also put in specs for shared libraries. Major players in the game are Sun, SCO and SAS Institute and Metrolink (of Metro-X accelerated server fame), but as far as I can see everyone can join, there are publicly available conformance test suites, etc. One rather obvious drawback from a FreeOS point of view is that they use Motif as part of the specfication. Find them at <http://www.metrolink.com/abi/index.html>. My, that was tricky to dig out. That particular acronym is heavily overused (including, BTW "Acquired Brain Injury" !!!!) -- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) 35327907