roberts@studguppy.lanl.gov (Doug Roberts) (05/02/89)
[Stuff about how tiresome Unix development environments can be] [Stuff about how wonderful LISPm's are] [Much more of this and it will begin to sound like the stupid grep vs. search wars going on in comp.software-eng] . . . Haven't we all heard this before? LISPMs are superior for developing LISP code. Granted. After all, that's what they were put on this world for. Unix boxes are superior general purpose machines. My point, (being made from the point of view of a former Symbolics bigot), the one I attempted to make about three follow-ups ago, is that each environment has its strengths and weaknesses. My own personal choice, after fours years of life in front of a 3600 was to switch to a Sun. And while I've cursed the sparsness of the LISP environment on the Sun as compared to my old 3600, my overall productivity has improved because of the other tools provided by Unix (grep, find, pipeing, awk, troff, tbl, eqn, me, tar, fsck & format [ever try to repair a bad block on a Symbolics disk? It can be done. It's not fun. For that matter, have you ever tried to boot your Symbolics diskless after its disk cratered?]...). On the optimistic side regarding LISP on the Unix boxes, I suspect that better development environments will be forthcoming. I've seen improvements in Lucid's product, and it wouldn't surprise me if other vendors (Envos, Franz, Symbolics??) didn't try to capture some of the Unix LISP market by providing more powerful development environments than currently exist. --Doug -- =============================================================== Douglas Roberts Los Alamos National Laboratory Box 1663, MS F-602 Los Alamos, New Mexico 87545 (505)667-4569 dzzr@lanl.gov ===============================================================
raymond@ptolemy.arc.nasa.gov (Eric A. Raymond) (05/03/89)
In article <ROBERTS.89May1220004@studguppy.lanl.gov> roberts@studguppy.lanl.gov (Doug Roberts) writes: >compared to my old 3600, my overall productivity has improved because >of the other tools provided by Unix (grep, find, pipeing, awk, troff, >tbl, eqn, me, tar, fsck & format [ever try to repair a bad block on a >Symbolics disk? It can be done. It's not fun. For that matter, have >you ever tried to boot your Symbolics diskless after its disk >cratered?]...). There is equivalent functionality for all of these on a LISPM. There just buried deep in the source code :-) I agree that some of the low level file system leaves much to be desired, but (to be realistic) I doubt it was really ever polished up for non-techy use. That what's field support is for. >On the optimistic side regarding LISP on the Unix boxes, I suspect >that better development environments will be forthcoming. Soon or later. Unfortunately it will be later and we've been hearing this for a LONG time. -- Eric A. Raymond (raymond@ptolemy.arc.nasa.gov) G7 C7 G7 G#7 G7 G+13 C7 GM7 Am7 Bm7 Bd7 Am7 C7 Do13 G7 C7 G7 D+13: Elmore James
mike@ists.ists.ca (Mike Clarkson) (05/03/89)
In article <1133@ptolemy.arc.nasa.gov>, raymond@ptolemy.arc.nasa.gov (Eric A. Raymond) writes: > I agree that some of the low level file system leaves much to be > desired, but (to be realistic) I doubt it was really ever polished up > for non-techy use. That what's field support is for. That's also why I could buy a Sun a year for what I was paying for hardware and software field support contracts. I would have had enough left over to have each of the new Suns supported as well. Mike. -- Mike Clarkson mike@ists.ists.ca Institute for Space and Terrestrial Science uunet!attcan!ists!mike York University, North York, Ontario, FORTRAN - just say no. CANADA M3J 1P3 +1 (416) 736-5611
buff@pravda.gatech.edu (Richard Billington) (05/05/89)
1) Fixing bad blocks is straight-forward now (lmfs:fix-file or si:fix-fep-file). 2) Annual maintenance on a MacIvory is around $500, the XL400 is similarly low but I don't have an exact quote, so I won't say. However, the Ivory technology has solved the big hardware maintenance cost problem for the future. 3) The biggest "problem" I'm aware of for Symbolics/Lispms in the ongoing "religious" wars is that for every 1000 unix boxes there's only 1 lispm. If the reverse were true, I suspect that unix wouldn't even be a recognizable word to most folks. It's really pretty primitive; a brilliant use of a couple of fairly simple ideas about file systems extended to be an operating system. A lispm was originally conceptualized as a program deveopment environment, not as a file system. The article that started this off and subsequent comments of "boy I'd miss piping, etc" only emphasize to me how broken people's concept of the right way to do things can become. 4) I've been a system administrator of both Symbolics Lispm's and unix workstations. System admin of unix is a holy terror by comparison (I've done both on and off for the last 6 years). 5) Please define "general purpose". A Symbolics runs C, Fortran, Pascal, Lisp, Flavors, and Prolog. Symbolics-based graphics systems are state-of-the-art (don't believe me? they've won an Emmy and have a massive presence at Siggraph). They now have a powerful hypertext system, an object-oriented, distributed, database system, several expert system construction tools, state-of-the-art CAD/CAM systems, Macsyma, X-windows, TeX and a TeX previewer that actually works (and has for a couple of years), etc. So, please, what does this mean: Lisp machines are not general purpose, but unix is?