Return to BSD News archive
Xref: sserve comp.unix.aix:48188 comp.unix.bsd:15410 comp.unix.pc-clone.32bit:7566 comp.unix.solaris:27508 comp.unix.unixware:14494 Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!yarrina.connect.com.au!werple.apana.org.au!daemon.apana.org.au!udel!gatech!howland.reston.ans.net!pipex!sunic!news.uni-c.dk!pd From: pd@kubism.ku.dk (Peter Dalgaard SFE) Newsgroups: comp.unix.aix,comp.unix.bsd,comp.unix.pc-clone.32bit,comp.unix.solaris,comp.unix.unixware Subject: Re: Unix for PC Date: 1 Dec 1994 16:22:04 GMT Organization: What? Me? Organized? Lines: 66 Distribution: inet Message-ID: <3bkt7c$b77@news.uni-c.dk> References: <199411210319.TAA18133@nic.cerf.net> <CHRISB.94Nov30100302@stork.cssc-syd.tansu.com.au> <3bihms$5pf@decaxp.harvard.edu> <palowodaD04HqA.LAJ@netcom.com> NNTP-Posting-Host: stat.kubism.ku.dk In <palowodaD04HqA.LAJ@netcom.com> palowoda@netcom.com (Bob Palowoda) writes: >What does Slackware cut a CD a week to keep up with these problems? Just >look at all the problems and unanswered questions that occur in >comp.unix.x.i386unix of people buying Slackware or the "The big Y's" forgot >how to spell it. Why do you think companies deploy source control and >patch revision control in the first place. With the many thousands of >contributors to Lunix their chaos to control has given proof to the old saying >"Too many cooks spoiled the pot". At admit FreeBSD has a little more >consistentcy to their releases. Now this is malicious slandering if ever I saw it. Have YOU seen the patch list for Solaris 2.3 lately? In fact, the kernel development of Linux is under tight control of only a handful of people, carefully coordinating patches and quickly fixing problems as they appear. For standard hardware, the problems have practically vanished by now (I'm running 1.1.52 -- yes, one of the socalled UNSTABLE kernels, "stable" ones have an even 2nd digit -- with an uptime of 36 days now and no sign of a glitch at all). If there is a problem, it is one of too many ingredients, rather than too many cooks -- there is an enormous variety of PC cards to cater for and umpteen ways in which they can step on each others toes. This is a problem for anybody trying to write OS's for PC's, not just the Linux operation. It is the happy users that are counted in thousands! Of course, the kernel is not the whole thing. There are also a number of important utilities, being developed along with the kernel. Some of these seem to have stabilized, whereas others are constantly being upgraded. This is the main reason for issuing new distributions of e.g. Slackware. Bear in mind that several of these packages don't come with the OS on commercial offerings. There, the sysadmin is the one that has to keep track of new releases of key PD programs: emacs, gcc, groff, TeX, ... Most of the problems with Linux and X come from newbies getting instantly promoted to sysadmins on a Unix system. Of course they get confused initially. Also, regarding X, the setup is admittedly a bit complicated, clocks and modelines and all that, but again this is mainly a matter based on PC hardware. Things are much easier if you only support only a limited set of framebuffer/monitor combinations. Once you get past this hurdle, things are no worse than on Suns (believe me, I've done both). In fact, Suns are sometimes more of a headache, because they insist on doing things in nonstandard ways. (Try getting imake to work properly, e.g.) As a sysadmin on both Suns and Linux systems, I can assure you that Slackware has been runnable at a level comparable to SunOS 4.1.x for at least the past year. Remaining problems are nowhere near comparable to what we are seeing currently with Solaris! (The argument of wanting to have somebody to sue when things go wrong is hard to argue against, but don't bash the software created by extremely talented and seriously working volunteers!) umount /dev/soapbox export FLAME=off -- O_ ---- Peter Dalgaard c/ /' --- Dept. of Biostatistics ( ) \( ) -- University of Copenhagen ~~~~~~~~~~ - (pd@kubism.ku.dk)