Path: utzoo!attcan!utgpu!utstat!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!unmvax!ncar!tank!shamash!nic.MR.NET!thor.acc.stolaf.edu!mike From: mike@thor.acc.stolaf.edu (Mike Haertel) Newsgroups: comp.unix.wizards Subject: Re: What kinds of things would you want in the GNU OS? Keywords: GNU OS features kernel Message-ID: <2273@thor.acc.stolaf.edu> Date: 24 May 89 17:35:35 GMT References: <106326@sun.Eng.Sun.COM> <10317@smoke.BRL.MIL> Reply-To: mike@stolaf.edu Organization: St. Olaf College, Northfield, MN Lines: 41 In article <10317@smoke.BRL.MIL> gwyn@brl.arpa (Doug Gwyn) writes: >In article <106326@sun.Eng.Sun.COM> bitbug (James Buster) writes: >>What kinds of things should be in the GNU Kernel? > >Why do you ask? Will this actually have an effect on the GNU kernel? > >My opinion is that the GNU kernel should either provide an exact >duplicate of a standard UNIX system interface, preferably SVR4, >or it should be a quantum leap forward in OS design. The latter >will not be achieved by piling "features" into it. No promises, but . . . The major goal of the GNU project has been compatibility with Berkeley UNIX, and then improvements. Now that POSIX is beginning to solidify, we are taking that into account. So I think it's fairly safe to say that we will be compatible with 4.3BSD (or perhaps 4.4, when it comes out), and that we will be compatible with POSIX. System V is a secondary concern, but we will probably try to be compatible with it insofar as it does not conflict with higher priorities. Compatibility with any of the above systems may be implemented at the library level rather than the system call level. We intend to support the usual protocols, like TCP/IP and NFS, but not necessarily in the kernel. I'm told that System V.4 is rather like two big rocks, stuck together with a small amount of glue. I don't see why that is `preferable.' I would tend to prefer pure Berkeley with a few concessions made (like a compatible tty driver) to ease the porting of System V stuff. In fact, I would tend to prefer pure-something-simpler-than-berkeley, with all concessions for compatibility at the library level or in user level servers. Some things I will agitate for are v9 style pipes and stream i/o, and putting system objects into the file system name space. I will also agitate against the creeping featurists, but I don't know if it will do much good. Even if the feeping creaturists win, I suppose those of us who care can spend a few months removing features to get a clean, small, system, and distribute it ourselves. -- Mike Haertel ``There's nothing remarkable about it. All one has to do is hit the right keys at the right time and the instrument plays itself.'' -- J. S. Bach