Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!saimiri.primate.wisc.edu!eng.ufl.edu!usenet.ufl.edu!zoyd!kem From: kem@cis.ufl.edu (Kelly Murray) Newsgroups: comp.os.386bsd.development Subject: Future Directions for BSD (was Re: What happened to these projects ?) Date: 27 Apr 1993 20:44:24 GMT Organization: University of Florida Lines: 63 Sender: kem@zoyd (Kelly Murray) Distribution: world Message-ID: <1rk5v8INN6vp@no-names.nerdc.ufl.edu> References: <1r855hINN56d@gap.caltech.edu> <hastyC5yJIs.Lqw@netcom.com> <1993Apr26.001822.7537@knobel.GUN.de> <hastyC62oIE.BFH@netcom.com> NNTP-Posting-Host: zoyd.prl.ufl.edu ------------------------------------------------------------------------------- Please allow me to chime in on my perspective for where BSD should be going. I'm a system programmer in the Parallel Research Lab here at Univ of Florida. I have a large amount of freedom to do what I find interesting. Let me also state that I am an entrepeneur, always looking for opportunities to find faster, cheaper, better ways to satisfy peoples needs that make good business sense. I have been using Linux happily at home, and in fact I have advocated it within the CIS department here, and we now have many Linux users. I have convinced my boss to buy me a 486 machine here at UF so I can do some development work with Linux (we have SUNS and SGI machines generally). However, this was before I really knew much about 386BSD. I really like the fact that it is not GNUified, and believe that this fact has the potential for BSD to have a much larger impact in general, as companies can be created such as BSDI, and hopefully many others that will compete with them, with real budgets for marketing the software, making it all the more visible, creating even more demand for free versions too. But if BSD is slowly going nowhere, I'm not going to jump on the wagon. I think one major reason that Linux has become popular is the SLS release. Most Linux users use it. It packages up much of the latest stuff in a single release that is easy to install, based on floppy disk distribution. BSD needs something like this. Someone needs to maintain the package as probably a full-time effort, putting it together and keeping it current, coming out with new releases when significant fixes or features are available. Keep it free, but also sell it as floppies, cart-tapes, CD-ROMS. I see this being different than the patch-kit (though I'm not all that familiar with it), because it is targeted for USERS, not developers. It is my impression that NetBSD effort is more oriented towards this approach, and that the 0.2 effort is less so, and as such I think the NetBSD folks are correct in taking the initiative to better serve the demands of PC Unix users. I agree with Amancio that it makes sense to identify components of the OS that people can sign up to work on. My impression of BSD is that is not truely a world-wide, distributed effort, but is very concentrated at UCB. Coming at this fresh from the outside, perhaps I don't have an adequate concern and appreciation for the Jolitz's and other UCB folks efforts, but I think that is because I am seeing Linux as more effective as a PC Unix. I also agree very much with Amancio, that BSD'ers need to actively advocate and promote using BSD. This is easier with a SLS-like package. I believe there is a significant difference with BSD and Linux in this regard. Many Linux'ers come from the DOS world, and so UNIX seems like such a dramatic step forward, wherease BSD folks have been around UNIX for so long they aren't as excited about it, and also are generally used to more state-of-the-art hardware (SGI's) and software (NextStep), such that 486's are ho-hum. I must admit that I have more interest in developing 90's technology than creating 70's technology for the masses. Perhaps it is possible to look foward to where BSD could go while maintaining compatibility with existing technology. This would put some excitement into the effort, while still creating useable software today. I think NextStep is a good model for this type of evolutionary approach. - Kelly Murray (kem@cis.ufl.edu)