Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.wildstar.net!news.ececs.uc.edu!news.kei.com!news.mathworks.com!news.pbi.net!news5.crl.com!nntp.crl.com!usenet From: "Jordan K. Hubbard" <jkh@FreeBSD.org> Newsgroups: comp.unix.bsd.freebsd.misc,comp.unix.bsd.bsdi.misc,comp.sys.sgi.misc Subject: Re: no such thing as a "general user community" Date: Sun, 30 Mar 1997 21:03:34 -0800 Organization: Walnut Creek CDROM Lines: 119 Message-ID: <333F45A6.41C67EA6@FreeBSD.org> References: <331BB7DD.28EC@net5.net> <5g9hjp$api@flea.best.net> <5gmb58$6jd$1@news.clinet.fi> <5gn3ig$83d@flea.best.net> <5goqrq$5ak$1@news.clinet.fi> <5hd29s$e7t@fido.asd.sgi.com> <333C1614.ABD@sgi01.grn.aera.com> <5hhv1k$jh9@fido.asd.sgi.com> <333E3530.794B@sgi01.grn.aera.com> <333EA3EF.41C67EA6@consys.com> <333EE416.ABD322C@FreeBSD.org> <5hn00k$dio@fido.asd.sgi.com> NNTP-Posting-Host: time.cdrom.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.01 (X11; I; FreeBSD 2.2-STABLE i386) Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:38118 comp.unix.bsd.bsdi.misc:6520 comp.sys.sgi.misc:29547 Larry McVoy wrote: > Jordan, your unhappiness with that Usenix paper has nothing to do with the > numbers. I just went through the numbers, which were certainly apples to > apples, slightly skewed in favor of FreeBSD (133Mhz P5 vs 120Mhz P5 for > Linux). Here's the summary of the i586 systems, FreeBSD vs Linux: Erm, I think you're sort of leaving off the high end Linux machine (I believe it was a P6/200) which you also tossed into this pot, comparing it openly with the FreeBSD machine's numbers on the P5/133. FWIW, I also wouldn't consider the comparison of the 133 Mhz P5 vs the P120 to be valid either since both the CPU and memory clocks are measurably slower in the latter case. But that only evades the principle issue, which is that I am NOT unhappy with the numbers or the fact that Linux or FreeBSD was the "winner" in your presentation (in order for "win" vs "lose" to even have any meaning for me, I'd first have to feel that the comparison was fair on both sides and I do not feel that way), nor do I think that the numbers themselves were "skewed" in any way. The numbers themselves were probably just fine. What I have a problem with is the *way the numbers were used* to prove a wholly irrelevant point about OS superiority. I was there at that presentation, Larry, and I listened to your both your words and your emphasis on them - you put up a big composite slide showing all the numbers, from P6 to 486, and then proceeded to focus almost entirely at the top-end P6 Linux numbers vs the "Big Iron" machines so that you could crow about how outstanding Linux's performance was. The other x86 OS numbers were on significantly wimpier hardware and all down in the noise somewhere, unworthy of serious mention, and one don't need to be Dale Carnegie to understand what the effects of such statistics are on an audience. Associate the OS you want to push with the really fast hardware and Bob's yer Uncle, that OS now looks very cool indeed! A presentation which would have certainly won a lot more points for proper procedure in my book would have involved some number (4? 5?) *identically configured* x86 machines, all running a different x86 OS variant, and compared straight on the merits of the OS software alone. Now THAT would have been an interesting talk, and seeing the hardware completely removed from the equation and the software tested across a wide range of operating conditions would have given all parties concerned some serious food for thought, but you evidently didn't want to go through all that trouble and instead you collected numbers off the net for entirely dissimilar hardware, destroying any chance for a truly productive comparison. Seeing you get up on stage and then try to SELL it as a productive comparison only added insult to injury, and that indeed is what gets my goat about the whole thing. The numbers themselves or the conclusions drawn don't bug me anywhere nearly as much as how you used them. > Which it is. In spute of your flames to contrary, my lack of support > for FreeBSD has nothing to do with the technology. I never said it did - what flames are we talking about? I don't flame you too terribly often, Larry, and the only times I can think of doing so involved my displeasure at your tactics in hyping Linux, not your lack of support for FreeBSD! Your support for FreeBSD would be very well received, don't get me wrong, but we hardly demand or expect it. All we ask is that you not let your enthusiasm get the better part of logic where evangelism for Linux (or whatever other OS you may someday choose) is concerned. > your version versus the other version? Wouldn't it be better if we were > all working on the same thing? Making one Unix better and stronger for > everyone? Yes. So why don't we take the older, more proven technology and run with it? That's the entire point of a lot of people's participation in the *BSD camps (and the fact that there is more than one BSD camp is a source of pain to us all and something we have worked long and hard, albeit unsuccessfully, to somehow fix) - we don't WANT to start over again with a totally new code base, source code control methodology and organizational structure. The CSRG had a very successful product in BSD (if you measure in terms of how much of it made it into commercial OSes or had a significant influence on their direction) and they left us a legacy that goes beyond just a collection of sources, it covers the build system, source tracking (we've switched from SCCS to CVS but the basic work methodology is the same), coding styles, documentation organization (in need of some cleanup but still servicable), etc and so forth. Why in the world we would want to do all this over again, especially when the Linux camp evidently doesn't even hold a lot of those things to be too terribly important (I'm thinking of things like SCM tools and a united will to use them, or a unified GNATs database for all kernel hackers), is really a mystery to me. Isn't there enough work to do in UNIX without reinventing the wheel? There are 15 years worth of work in BSD and we've PROVEN that it's highly portable (re: NetBSD), prove that it can perform very well under load (re: FreeBSD and cdrom.com/Yahoo/etc) and proven that can be a commercial success (re: BSDI). Geeze, what more do you want, an endorsement from some famous basketball player? Would that make us cool and hip enough to get on board with? :-) Lest you next say that we're clinging to a stale code base, let me also say that our feature set is hardly frozen, and a number of really good networking people (just to name one category of development) are busily taking BSD's networking technology forward in a number of new directions. Popularity? More popularity would be nice, yes, but we're getting there too - I see no signs of FreeBSD's growth slacking off at all, despite what some of you doomsayers occasionally predict (and yes, I'm collecting registration & download statistics for FreeBSD now so I don't have to say "I think" when I say it's growing), and that too will address itself in time - it's just a matter of getting enough users to reach a certain momentum, and we're still a year behind Linux in terms of our actual time-on-the-market. Give us a chance! ;) > And if you disagree and want to continue to work on your version, go > for it. Perhaps you can do wonderful things in your OS and that will > be great. But try and understand that I am actually trying to make sure > that those wonderful things are relevant. If the world is 100% NT, Me too. :-) I think we share a lot of the same goals, Larry, we've just chosen different ways of reaching them. -- - Jordan Hubbard FreeBSD core team / Walnut Creek CDROM.