jwz@lucid.com (Jamie Zawinski) (11/29/90)
Bjorn Victor wrote: > And, to lead into the other subject, it's a real nice hacking machine. > I can spend hours fixing a subtle bug, or re-implementing a nice > feature from somewhere else, and it's FUN! On the other hand, it's > fun hacking Emacs lisp too (e.g. re-implementing Zmacs features), and > on the third hand, I really should do something completely different > -- my research. But Zmacs is so much better than GNU Emacs in so many ways... even down to the datastructures! GNUmacs is character-oriented rather than line-oriented, which makes a lot of things next to impossible to do. Lines don't have plists; you can't elide arbitrary sections of text the way VisiDoc does; the first insert after a move in any file larger than ten megs takes up to half a minute (because GNUmacs has to move the one and only one "insertion hole"). Minor modes in GNUmacs are almost impossible to implement correctly. I think TAGS tables suck. Both conceptually and in the current implementation. I could go on and on... Also all of the GNU Emacs mail readers are substantially more lame. Hans Chalupsky: >- development environments (such as various fancy lisp-modes for GNU emacs) > are (in my opinion) better then the TI-Explorer environment (there are still > a few things missing such as window based debuggers and inspectors, but you > can get approximations of those in the Allegro composer or Sun's SPE) Pale imitations at best, and really slow from what I've seen... The Lispms have the state-of-the-art in development environments, even though they're going the way of the dinosaurs. But we'll fix that soon :-) Bjorn: > I think Emacs will eventually have most of Zmacs' "graphic" features (e.g. > multiple fonts), but some of the other bugs are very hard to get out. And it > will always live with 8-bit characters, which is a loss from the 12-bit LISPM > characters. You can't both have "clean" ISO 8859-1 and a Meta key, for > instance. I'm not convinced: just because you can't have 12-bit characters in a file doesn't mean you can't have 12-bit input. If input was event-based instead of character based, then under X you could easily have seperate keybindings for left-shift-cokebottle and right-shift-cokebottle... Do any of you have Explorers at home? I've often considered buying a used one to play with, but I'm a bit concerned about power/temperature requirements, and availability of spare parts. Anyone have any comments? -- Jamie, considering getting CDRNEXT on my license plate.
miller@cs.rochester.edu (11/29/90)
Well, yes, I do have an Explorer at home, though it wasn't because I bought
it. We unplugged 6 of them (3 IIs and 3 Is) and couldn't find any buyers, so
they let me take a II home and scavenge. Right now it's 8meg, 2 bricks &
tape, and I'm quite happy with it, thank you :-). Alas, I'm only running 4.1
as we dropped software maintainance while 5.something was the most recent
(only for the microexploders).
Noise is a slight problem, tempertaure not at all, but then I have
electric heat anyway :-). The only thing I really need to do with it
"soon" is run a serial line between it and my mac - that way I can xfer
files a bit easier than dump to tape, walk to work, download on my
modem, and I'd love to use Zmacs instead of any of the much less good
Mac editors available, esp. for Lisp. (I'm developing stuff under MACL).
As for spares, well, I can just strip boards out of our other "surplus"
machines as much as necessary. It's how we're keeping our 36xxs running too
(though perhaps not much longer :-).
If it matters, I use a UX400s at work which I'll only give up for a UX1200s.
BTW: if anyone is interested in purchasing an Explorer I or II, make offers
to bukys@cs.rochester.edu, I'm sure he'd love to turn them into cash.
CDRNEXT? How about CDDADR :-)
--
----
Brad MillerU. Rochester Comp Sci Dept.
miller@cs.rochester.edu {...allegra!rochester!miller}Bjorn.Victor@docs.uu.se (Bjorn Victor) (11/29/90)
>But Zmacs is so much better than GNU Emacs in so many ways... even down to >the datastructures! GNUmacs is character-oriented rather than line-oriented, >which makes a lot of things next to impossible to do. Lines don't have >plists; you can't elide arbitrary sections of text the way VisiDoc does; >the first insert after a move in any file larger than ten megs takes up to >half a minute (because GNUmacs has to move the one and only one "insertion >hole"). Minor modes in GNUmacs are almost impossible to implement correctly. >I think TAGS tables suck. Both conceptually and in the current implementation. >I could go on and on... Yes, I agree on most of this, although I don't quite know what you mean with minor modes being "almost impossible" to implement correctly? The data structures is a big problem, as well as that C-coded "lisp functions" call other C-coded "lisp functions" directly, making redefinition in Lisp impossible. To redefine, rebuild... On the other hand Emacs is much faster than Zmacs most of the time, even though it is byte-coded and has a lousy compiler. Also, I think it's fun to hack in systems that are powerful enough to be hacker-friendly, but feature-lacking/buggy, like Emacs or Zmacs or Explorers (or TECO, but that was quite a while ago now). >Also all of the GNU Emacs mail readers are substantially more lame. Yes, but they are much faster (and I guess the machines I run Emacs on have faster file systems). I've converted from ZMAIL to RMAIL because it's faster and the file system is better. Also, my 1.5MB BABYL file broke on the Explorer (it was suddenly 5MB, where the extra space was epsilons, so Zmacs couldn't read it). >Hans Chalupsky: >>- development environments (such as various fancy lisp-modes for GNU emacs) .... > >Pale imitations at best, and really slow from what I've seen... The >Lispms have the state-of-the-art in development environments, even though >they're going the way of the dinosaurs. But we'll fix that soon :-) Tell us more, you DO work on a serious Lisp company nowadays... >Bjorn: >> I think Emacs will eventually have most of Zmacs' "graphic" features (e.g. >> multiple fonts), but some of the other bugs are very hard to get out. And it >> will always live with 8-bit characters, which is a loss from the 12-bit LISPM >> characters. You can't both have "clean" ISO 8859-1 and a Meta key, for >> instance. > >I'm not convinced: just because you can't have 12-bit characters in a file >doesn't mean you can't have 12-bit input. If input was event-based instead of >character based, then under X you could easily have seperate keybindings for >left-shift-cokebottle and right-shift-cokebottle... OK, true, but do YOU want to hack Emacs to handle that? On the other hand, I think/dream that the world will *eventually* (in 5-10 years or so) go 16-bit, to cope with larger character sets than ISO 8859-n. But then all 16 bits will be char-code, and as usual no bits will be left for Control/Meta/Hyper/Super/Command/Alt/Whatever... Maybe in 20 years we'll get it again? >Do any of you have Explorers at home? I've often considered buying a used one >to play with, but I'm a bit concerned about power/temperature requirements, >and availability of spare parts. Anyone have any comments? I've been thinking about it, but I think noise and spare parts are the preventive issues. > -- Jamie, considering getting CDRNEXT on my license plate. I like it! I'll consider CDRNIL (we can only have 6 chars in Sweden, and my car isn't very fast). -- Bjorn Ps. Hey, I think this list came alive again!?
Bjorn.Victor@docs.uu.se (Bjorn Victor) (11/29/90)
>If it matters, I use a UX400s at work which I'll only give up for a UX1200s. Could you tell me more about these Symbolics cards? I haven't read anything about them, all I think I know is that they are based on the Ivory chip, which I read about when they were developing it. >BTW: if anyone is interested in purchasing an Explorer I or II, make offers >to bukys@cs.rochester.edu, I'm sure he'd love to turn them into cash. Hmmm... thanks for the info! -- Bjorn
miller@cs.rochester.edu (11/30/90)
Basically the XL400/XL1200 is a VMEbus machine. Symbolics will sell you just
the CPU-board (which includes 20meg of memory, the FEP, etc.) to plug into
your sun. User access is via X, plus life support routines so the SUN (say
a 4/330, but anything with a VMEbus should work) supplies the disk, the
console, the network connection, and the UX piggybacks on all that. From
a user point of view, a screen comes up on your machine (any x server, not
necessarily the host machine - the host just needs to supply disk, and
a slot) which looks just like the console screen of a stand-alone symbolics,
has the same windows, etc. The keyboard is mapped into the symbolics character
set, (configurable to your tastes) and you are running!.
Note that you can plug in >1 board into a host, if you want to say, run
a centralized 4/490 with 6 or so UX boards in it for remote use. I find
running on the local machine only marginally faster than running remotely
on a Sparc 1. (in the previous relase the difference was much greater).
Network traffic is higher in the remote case, obviously.
--
----
Brad MillerU. Rochester Comp Sci Dept.
miller@cs.rochester.edu {...allegra!rochester!miller}Moon@STONY-BROOK.SCRC.SYMBOLICS.COM (David A. Moon) (11/30/90)
Date: Thu, 29 Nov 1990 13:53 EST
From: rochester!miller@rutgers.edu
....
the SUN (say a 4/330, but anything with a VMEbus should work) ....
Yes, that's what you'd think if you weren't an expert on the VMEbus and
SUN's many VMEbus implementations. The reality is different. It's
really quite appalling.
I say this not to criticize SUN, but just as a warning not to count on any
configuration working that Symbolics does not say is a supported configuration.
Re the subject line of this message: I don't know if people still use
Explorers, but it certainly appears that they don't still talk about them!pf@islington-terrace.csc.ti.com (Paul Fuqua) (12/01/90)
Date: Wednesday, November 28, 1990 4:25pm (CST)
From: agate!shelby!neon!lucid.com!jwz at apple.com (Jamie Zawinski)
Subject: Re: Do people still use Explorers?
I think TAGS tables suck. Both conceptually and in the current implementation.
I could go on and on...
A lot of stuff you get for free from the integration of tools in a lispm
(like m-. without tag-tables) you often have to do in a roundabout way
on a Unix Lisp.
One of the best examples is something called ilisp, a gnuemacs package
built on the CMU comint stuff that provides a reasonable Lisp listener
and editor commands, and calls on the inferior Lisp for things like
arglists, source-file-names, and the like. It's not a bad imitation of
the Zmacs/listener combination, but it has to work hard to do it.
Also all of the GNU Emacs mail readers are substantially more lame.
I agree. Even though I've made rmail behave much like the Explorer mail
reader (and both much like Babyl), I find rmail much more fragile --
it frequently duplicates the whole header while reforming if it sees
something it doesn't like.
Pale imitations at best, and really slow from what I've seen... The
Lispms have the state-of-the-art in development environments, even though
they're going the way of the dinosaurs.
A group I've worked with looked at SPE, Composer, and Harlequin, and
while Harlequin looks by far the best, none of them have (to my mind)
decent debuggers. I don't care about fancy windows and displays -- I
don't use the lispm window-debugger at all -- but I do want to be able
to grab and manipulate the args and locals, force returns from stack
frames, restart calls with new arguments, and look at the disassembled
code. I haven't found any Unix Lisp that lets me do all of them.
Then there's speed -- my 25 MHz, 24 MB Explorer 2 still beats a
neighbor's SparcStation 330 (32 MB, Harlequin, native CLOS) on a
cut-down version of a simulation I've been writing. The 330 is about
80% of the Explorer for this example, but the Explorer processor is four
years old.
And it
> will always live with 8-bit characters, which is a loss from the 12-bit LISPM
> characters.
It is forever annoying that I can't do c-m-rubout (Backward Kill Sexp)
in gnuemacs -- there's no c-rubout to put a meta modifier on.
Do any of you have Explorers at home? I've often considered buying a used one
to play with, but I'm a bit concerned about power/temperature requirements,
and availability of spare parts. Anyone have any comments?
Power and temperature should be okay, since it was designed for an
office environment. We ran two (five bricks, total) off a 20-amp
circuit in a conference room, but I wouldn't recommend it (we popped the
breaker when someone else fired up a NeXT on the same circuit).
I've thought about buying the one on my desk when it's replaced. Except
for the processor and the second brick, it's 5+ years old, so it should
be depreciated down to nothing. On the other hand, it's been years
since the last hardware failure, so maybe everything's about to break.
Paul Fuqua pf@csc.ti.com, ti-csl!pf
Texas Instruments Computer Science Center, Dallas, Texas
"I f***ed up." "Yes, but you gave it 100% effort." -- Mystic Pizza