Return to BSD News archive
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!bunyip.cc.uq.oz.au!munnari.oz.au!news.Hawaii.Edu!ames!hookup!news.mathworks.com!europa.eng.gtefsd.com!howland.reston.ans.net!gatech!rutgers!uwvax!grilled.cs.wisc.edu!jcargill From: jcargill@grilled.cs.wisc.edu (Jon Cargille) Newsgroups: comp.os.386bsd.misc Subject: The Real reasons the *BSDs should cooperate, if not merge. Date: 27 Oct 1994 22:57:39 GMT Organization: U of Wisconsin CS Dept Lines: 66 Message-ID: <38pb93$ls9@spool.cs.wisc.edu> NNTP-Posting-Host: grilled.cs.wisc.edu [...I'm sending this letter to the core groups of both FreeBSD and NetBSD, as well as posting it in the USENET groups. Sorry for the length of this...] Up until now, the fact that there is more than one free version of BSD for the i386 has been at most an inconvenience: it leads to extra work for those who make binary packages available on the net, since they have to make two versions; it's a source of confusion for new users, who don't know what to install; it causes some delay for end-users in terms of getting new features in their hands, since they might be implemented first in the "other" camp, and take time to migrate over. Inconvenience aside, however, the split hasn't been a REAL problem for many people. In the future, though, I see it becoming a true Achilles heel for the user community. The real problem is just beginning to rear it's head as *BSD begins to gain more visibility and legitimacy in the marketplace. We're seeing the first signs of this now with a few commercial packages that are being made available for *BSD. For example, there is now a version of Accelerated-X from X Inside for FreeBSD. There is now a beta version of AFS available from Transarc (courtesy of the folks at MIT who did the work) for NetBSD. I think it's great to have commercial outfits take an interest in releasing their products for *BSD. If we're really lucky, they'll release binaries for both. However, they're unlikely to want to do the extra work to produce releases for two close-to-identical OSes. The fact that there's no easy way for them to release binaries for both versions of *BSD may prevent some from releasing any binaries at all. The other possibility that they'll release their product for only one of the *BSDs. The two examples I mention above have chosen this path, and I think this is what most commercial outfits will do. Unfortunately, this means that it's the users that lose: if you want to run package X, you must run FreeBSD; however, if you want to run package Y, you must run NetBSD. If you want to run X & Y, you simply lose. The probability that you will want to run an incompatible X & Y increases exponentially with the number of binary packages that are released for only one version of *BSD. Call to Arms: At the *VERY* minimum, I think it would behoove the *BSDers to move to a single ABI, so that each OS could run binaries compiled for the other. Personally, I still think everyone would benefit if the two groups could get along well enough to implement either a partial (kernel-only), or more substantial (whole-world) merge. Any comments from the Peanut Gallery? Cheers, Jon, well asbestos-ized. -- -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-. Jon Cargille jcargill@cs.wisc.edu Want your .sig compressed? Reasonable rates and fast turnaround. Call today!