Date: Thu, 01 Feb 2001 02:22:01 +0100 From: Roelof Osinga <roelof@nisser.com> To: Mike Meyer <mwm@mired.org> Cc: "Albert D. Cahalan" <acahalan@cs.uml.edu>, questions@FreeBSD.ORG Subject: Re: OT: non-Unix history (Was: FreeBSD vs linux) Message-ID: <3A78BA39.8A14F8F@nisser.com> References: <14957.31196.939559.889627@guru.mired.org> <3A6F43F7.E43C6CA0@nisser.com> <14959.23870.728403.859934@guru.mired.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Mike Meyer wrote: > > ... > There are *lots* of potential reasons, many of them good ones. The > issue about machine speed shows up in the infamous "Worse is better" > paper, and I talk about this particular case in my "Good enough is > best" paper <URL: http://www.mired.org/home/mwm/good-enough.html > > (which provides pointers to "Worse is better" as well as covering the > salient features). I've started glimpsing part of that, or rather a reference in that paper (The Rise of ``Worse is Better'') and I must say that what is said in that is more or less ancient history. By which I mean predating by far that article. It is commonly known as Keep It Simple, Stupid! or KISS for short. Something I, like any engineer, fully subscribe. Maybe one of these days I'll read it fully. Until then I would like to remark that introducing new terminology is in direct conflict with the premise of KISS. New Jersey approach? Sheesh! It's a village type clash. Like the whole looser-ing problem. To me it depends one the drawing of the system boundary. At what are you looking? The whole solution or just one aspect? The difference between wanting the procedure perfect versus the program perfect. The latter does not necessarily depend on the former. Hm. It's *up to* and maybe including half that are worse (or better) than median. Some are likely to be median and therefore not worse :). I don't quite agree with heaping CL and scheme together. Scheme is more aking to LISP than CL. Originally anyway. Like comparing C++ as CL with C as Scheme which were surprisingly used interchangingly. It's also too simplistic. I spent several years working on an implementation (and design, naturally) of a functional language. Yet I chose to make it strict, not lazy, define it recursively (i.e. in a strict language) and make it pragmatic. So though FL in itself is the pinnacle of MITness that, in itself, does not preclude KISSiness <g>. I called it TLC for Typed Lambda Calculus, later renamed to Tender Loving Care by one Hal Hildebrand. Whom I had send the source since he asked for it. 'T was Smalltalk source with a C VM, btw. One twist I observed in my hastily glimpsing of your piece is that you seem to skip over the fact that what 'worse is better' is saying, namely that KISS will win the battle, is precisely that. Meaning to ultimately got 'something' into the most hands. Does refrasing constitute a caricature? Then again, as someone who tries to adhere to KISS I'm biased. To me they both say more or less the same thing. The differences being so minor a detail as to mean that bothering about those conflicts with the whole idea of KISS ;). > ... > to port it to each variant. Since VMS - and later NT - were usually a > larger market than any single Unix vendor, even if it wasn't as big as > all of them put together, it got preference. Which, of course, is why MVS or even OS/400 still rule the day. No? Doesn't work that way, I'm afraid. Also, define 'usually'. Are we talking early '93 'usually' or early '00 'usually'-ness? I'm asking because in '93 UNIX ruled whereas in '00 UNIX still ruled but the trade rags said 't was NT that ruled <g>. So define market. If UNIX is a specialisation of Multics and VMS a specialisation of UNIX, then NT is but another UNIX <g>. Granted, with a whole different API/ABI. So what else is new? <g,d&r> Roelof -- @ @ http://BeerIsBitter.com/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3A78BA39.8A14F8F>