I wish I'd never met Lisp
David Steuber
especially quickly if you don't like rants that are ill thought out
and clearly the ravings of a lunatic.
I wish I'd never met Lisp. It has ruined me for all other
programming languages. C and Java now look really ugly to me. XML
is clearly someone's really bad joke at the world. Jelly is a new
player in the Java/XML conspiracy. I'm not sure how to con someone
into paying me to hack on Lisp.
Since I've gotten my Mac, I've been ruined for all other desktop
systems. Sadly, there are a limited number of Lisp systems that work
on it well. The only one I've been able to compile myself is SBCL.
OpenMCL should be better to use because it has a Cocoa bridge and
does threads. I haven't been able to build it yet though (I have an
"old" binary build from Darwin Ports). OpenMCL is not really ready
for prime time yet because any Cocoa app built with it apparently
will only work under the exact revision of OS X that it was built
for.
Lispworks costs $999 and MCL runs $750. I didn't look at Franz.
I've had enough sticker shock for one day. I'm not buying a
commercial Lisp soon, but I have downloaded the Lispworks Personal
Edition to play with to see how I like that.
Emacs + SLIME looks really primitive next to Xcode. Xcode has all
sorts of cool features to (once it is finally running). Xcode will
take care of all the gory details of making NIB files and building a
.app bundle for your application. The only catch is that you have to
program in Objective-C or Java (preferably the latter). Did I
mention that Lisp has ruined me for all other programming languages?
Emacs + SLIME + SBCL take up less memory than Xcode. The startup
time is shorter. The programming experience is far more
interactive. Yes, it looks primitive by comparison. In reality, it
seems a bit more sophisticated. The problem is, I am just not loving
Emacs so much (mostly due to key chords). Yes, I use it for mail and
gnews. That doesn't mean I like it. There is no Cocoa bridge and no
.app packager. There is no integration to Xcode and its tools at all.
I think I would like to have that integration. I would like to
program in Lisp that way. I just don't have $999 to spend on it.
Besides. I think that a worthy free Lisp implementation and dev
environment for OS X is very useful for Lisp's future as a perhaps
not so dead language. Obligatory Miracle Max quote goes here.
So I've got a wish list. It's not a long list. Call it a short
list.
* SBCL should support threads on OS X using the same programmer level
API as on Linux x86. If someone points me to where to look at
hacking that in, I'd be happy to help.
* SBCL should have a Cocoa bridge and Carbon wrappers for OS X.
Again, I'm willing to work on this.
* There should be a nicer IDE for Lisp that fits the OS X Aqua
interface guidelines. Fewer key chords than Emacs has would be
nice. SLIME like functionality is a must. The IDE should be part
of the Lisp image. All forms of RPC suck. A threaded Lisp would
help.
* The Lisp IDE should be able to leverage Xcode for Cocoa, Carbon,
Quicktime, and other documentation. App bundles and also
frameworks should be things that you can make.
It's time to make Lisp mainstream.
--
An ideal world is left as an excercise to the reader.
--- Paul Graham, On Lisp 8.1
Kenny Tilton
David Steuber wrote:
> I wish I'd never met Lisp. It has ruined me for all other
> programming languages.
...and...
> Lispworks costs $999 and MCL runs $750.
Sit in the lotus position until the light goes on. Hint: I starved for
months to afford a $1600 Apple frickin II in 1978 so I could do integer
basic in a whopping 16K of RAM, and I was happy to have it.
If the light does not go on, I suggest you arrange for another
twenty-thousand man-hours priced at ten cents each so you can get
CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial lisps.
It is /so/ important that our tools be free, no matter what the cost.
<sigh> C'mon, Lispniks are supposed to be /intelligent/.
Am I being obscure? Try this: if you cannot find $750 for a commercial
Lisp, you either are not really a programmer or you do not really like
Lisp. Your choice.
Hope you like rants. :)
kenny
--
Home? http://tilton-technology.com
Cells? http://www.common-lisp.net/project/cells/
Cello? http://www.common-lisp.net/project/cello/
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Your Project Here! http://alu.cliki.net/Industry%20Application
mikel
If you like. I'll help you get OpenMCL built from the latest sources on
your system.
It's true that Cocoa apps built with OpenMCL 0.14.x run only on pretty
much exactly the version of OSX that built them. However:
1. Gary and Randall are working on improvements to the Cocoa bridge to
fix that problem.
2. Carbon apps built with OpenMCL have no such problem.
3. Bosco 0.6 (see http://www.common-lisp.net/project/clotho/) builds
both Cocoa and Carbon apps. Bosco 0.6 apps have SLIME built into them so
you can connect to the running app from Emacs.
4. I'm working on a short tutorial with example code that shows
step-by-step how to build both Carbon and Cocoa apps on Bosco.
In addition, the Clotho project is intended to produce an IDE for
OpenMCL on OSX. Clotho 0.2 suffers from the OS-version problem, but it
has a built-in GUI listener and inspector, and an extensible editor
written in Common Lisp. Clotho 0.3 will be built on Bosco 0.6, and will
be Carbon-based, and so not tied to a specific OSX version. Paul Lathrop
has contributed code to autobuild an application bundle in Clotho 0.3,
and I've added features to the Bosco 0.6 build system to build
applications into the right part of the bundle. It should not be hard to
add features that do the right things with Info.plist and other resource
files, and Lisp code can use nibfiles produced by Interface Builder just
fine. The tutorial will cover that topic as well.
One way to address the problems you've noted is to join the Clotho
project and contribute some code.
I'd be happy to make these same things happen for SBCL, but not until
the OpenMCL work is a good deal farther along.
Wade Humeniuk
>
> Since I've gotten my Mac, I've been ruined for all other desktop
> systems.
How much did you pay for your Mac?
Wade
Luke Gorrie
> If the light does not go on, I suggest you arrange for another
> twenty-thousand man-hours priced at ten cents each so you can get
> CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial
> lisps.
Apologies if this has been hashed out before, but what are the
features that make the commercial Lisps so much better?
I'm particularly interested in IDE features that are applicable to an
Emacs-like environment. I'm also curious to know what Lisp development
environment the long-time lispers in this group use, and why. For
instance, Joe Marshall, how do you hack your Lisp nowadays?
Cheers,
Luke (not soliciting "rants")
Edward Jason Riedy
- Am I being obscure? Try this: if you cannot find $750 for a commercial
- Lisp, you either are not really a programmer or you do not really like
- Lisp. Your choice.
Or you develop software that's expected to run on 100-9000
processors and cannot afford to force your clients / users to
afford $750 per processor.
Particular machine in question: seaborg.nersc.gov . I wish
they could magically get the _one_ compiled Lisp that runs on
that platform, but it ain't going to happen politically. And
I need to run on the remaining T3Es and think real hard about
the SV1. No support there at all.
"We don't use Lisp" even if it would make life better. Can't
beat those politics _and_ cost >>$750 per proc (or node, I
don't recall). (And I've never managed to get a free CL-ish
impl. to work in 64-bit mode in AIX 5.2, including clisp . Oh
well. No CL for me.)
Jason
--
Brian Downing
Edward Jason Riedy <e...@cs.berkeley.edu> wrote:
> Or you develop software that's expected to run on 100-9000
> processors and cannot afford to force your clients / users to
> afford $750 per processor.
>
> Particular machine in question: seaborg.nersc.gov . I wish
> they could magically get the _one_ compiled Lisp that runs on
> that platform, but it ain't going to happen politically. And
> I need to run on the remaining T3Es and think real hard about
> the SV1. No support there at all.
>
> "We don't use Lisp" even if it would make life better. Can't
> beat those politics _and_ cost >>$750 per proc (or node, I
> don't recall). (And I've never managed to get a free CL-ish
> impl. to work in 64-bit mode in AIX 5.2, including clisp . Oh
> well. No CL for me.)
I like free software a lot, and use free Lisps almost exclusively.
Unfortunatly this argument doesn't hold up. LispWorks will let you
deliver applications with no distribution charge for the runtime. The
limitations? As far as I know, it disables COMPILE-FILE. That's it.
If you're making any money at all off your software (or if it saves you
an appreciable amount of time) that's a amazingly good deal.
-bcd
--
*** Brian Downing <bdowning at lavos dot net>
Erann Gat
Luke Gorrie <lu...@bluetail.com> wrote:
> Kenny Tilton <kti...@nyc.rr.com> writes:
>
> > If the light does not go on, I suggest you arrange for another
> > twenty-thousand man-hours priced at ten cents each so you can get
> > CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial
> > lisps.
>
> Apologies if this has been hashed out before, but what are the
> features that make the commercial Lisps so much better?
FRED. (That's the editor in MCL.)
It's like EMACS, but it's written in (and more to the point, hackable
in) Common Lisp. So I don't have to learn a new dialect of Lisp to
program my editor.
Fred is the one feature that has kept me a loyal MCL user for nearly
twenty years. I started with the first release of Coral Common Lisp and
I've never looked back.
E.
Christian Lynbech
David> I'm not sure how to con someone into paying me to hack on Lisp.
I am working for Ericsson which, as any big corporation, is rather
un-excited about lisp.
But I am constantly thinking/planning/dreaming/working on slipping
lisp in somewhere. Even if that has little chance of actually
succeeding, the activity itself helps keeping my sanity up in this mad
corporate world.
As somebody famous once said: "We may be lying in the gutter, but we
are looking at the stars."
------------------------+-----------------------------------------------------
Christian Lynbech | christian #\@ defun #\. dk
------------------------+-----------------------------------------------------
Hit the philistines three times over the head with the Elisp reference manual.
- pet...@hal.com (Michael A. Petonic)
Andreas Krey
...
> I am working for Ericsson which, as any big corporation, is rather
> un-excited about lisp.
>
> But I am constantly thinking/planning/dreaming/working on slipping
> lisp in somewhere.
Did so by writing an interpreter used in some small tools. :-)
> Even if that has little chance of actually
> succeeding, the activity itself helps keeping my sanity up in this mad
> corporate world.
Doesn't work for me. I would need some un-scheme with some features
(re builtin types) and would like to implement it but can't rip
off the time anywhere. :-(
Andreas
--
np: 4'33
Espen Vestre
> Fred is the one feature that has kept me a loyal MCL user for nearly
> twenty years. I started with the first release of Coral Common Lisp and
> I've never looked back.
Yes, FRED is really a wonder! I think it was the tremedously ease
of extending FRED into anything you like that really opened my eyes
to the usefulness of Lisp for general-purpose programming (Hmm! I
didn't think of that when I put down a few lines of "my road to
lisp" recently...).
--
(espen)
Espen Vestre
> Unfortunatly this argument doesn't hold up. LispWorks will let you
> deliver applications with no distribution charge for the runtime.
Last time I checked, they still charged for Unix (not linux or OS X)
runtimes (but $750 for e.g. a Solaris run time license is still a
bargain).
--
(espen)
Espen Vestre
> Apologies if this has been hashed out before, but what are the
> features that make the commercial Lisps so much better?
I do some cross-platform GUI development, and LispWorks's CAPI
saves me a lot of work. I also use the graphical debugging/inspection
tools frequently, especially function call graphs and the generic
function and class browsers.
--
(espen)
Alain Picard
> Apologies if this has been hashed out before, but what are the
> features that make the commercial Lisps so much better?
I guess the integrated GUI builders would have to be the
killer feature. After that, things like the inspector,
graphical debugger and class browsers are gravy. Then
come the nice but not strictly required things like profilers
and defsystem browsers.
Well, that's what I think is way superior about, say, LispWorks
than CMUCL. Also, LW comes with CommonSQL (which is now probably
mostly replacable with UncommonSQL, but that wasn't true a couple
of years ago) and CORBA orb, still absent from the open source world.
Of course, CMUCL kick's LW's butt in any sort of numerical computation,
and it doesn't have the ridiculously low value of ARRAY-DIMENSION-LIMIT. :-(
Finally, another incredibly powerful feature is: support.
Tim Bradshaw
>
> Am I being obscure? Try this: if you cannot find $750 for a commercial
> Lisp, you either are not really a programmer or you do not really like
> Lisp. Your choice.
>
SHUT UP KENNY! You must know that you are NOT ALLOWED to be
economically rational in public. They have *no idea* how much things
cost and it's *critically* important to keep it that way if our
masterplan is to succeed.
For God's sake, man, you'll be telling them where we get the
helicopters next.
--tim
Rob Warnock
+---------------
| For God's sake, man, you'll be telling them where we get the
| helicopters next.
What, you mean like these?
<URL:http://www.helispot.com/photos/03269.html>
<URL:http://www.bellhelicopter.textron.com/en/aircraft/military/
bellAH-1Z.cfm>
<URL:http://www.rotaryaction.com/images/clearapd.jpg>
-Rob
-----
Rob Warnock <rp...@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
Russell McManus
> And Kenny Tilton writes:
> - Am I being obscure? Try this: if you cannot find $750 for a commercial
> - Lisp, you either are not really a programmer or you do not really like
> - Lisp. Your choice.
>
> Or you develop software that's expected to run on 100-9000
> processors and cannot afford to force your clients / users to afford
> $750 per processor.
Lispworks licensing does not work this way. Please check your facts
before posting mis-information.
-russ
Paolo Amoroso
> player in the Java/XML conspiracy. I'm not sure how to con someone
> into paying me to hack on Lisp.
You may start asking around among your friends, relatives and
colleagues who are not computer savvy. Some of them probably need a
small application or tool. And, provided it works, they probably
don't care which language it's written in.
In short, start with grocery store size businesses. You may not
become rich quick, but this may bootstrap a for pay Lisp coding
activity.
Paolo
--
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Recommended Common Lisp libraries/tools (Google for info on each):
- ASDF/ASDF-INSTALL: system building/installation
- CL-PPCRE: regular expressions
- UFFI: Foreign Function Interface
Marco Antoniotti
vendors will clearly say here on CLL that that is what they will charge you.
I am not saying that it will be cheap. I am saying that my experience
with the vendors is that they will work with you on something like this.
Cheers
--
Marco
André Thieme
> Tim Bradshaw <tfb+g...@tfeb.org> wrote:
> +---------------
> | For God's sake, man, you'll be telling them where we get the
> | helicopters next.
> +---------------
>
> What, you mean like these?
>
> <URL:http://www.helispot.com/photos/03269.html>
> <URL:http://www.bellhelicopter.textron.com/en/aircraft/military/
> bellAH-1Z.cfm>
> <URL:http://www.rotaryaction.com/images/clearapd.jpg>
>
Cool, I just ordered one! ;-)
André
--
Kenny Tilton
Luke Gorrie wrote:
> Kenny Tilton <kti...@nyc.rr.com> writes:
>
>
>>If the light does not go on, I suggest you arrange for another
>>twenty-thousand man-hours priced at ten cents each so you can get
>>CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial
>>lisps.
>
>
> Apologies if this has been hashed out before, but what are the
> features that make the commercial Lisps so much better?
In one word: complete integration.
Editor, listener, debugger, inspector, various browsers all offered by
one process which sees it all. So I can click on a stack frame in which
function X gets invoked and hit CONTROL-ALT-. (jump to definition in
source) and be investigating the source in about two seconds. And this
pattern works no matter where I am and no matter which other
navigational tool I want to kick off.
One measure of the importance to this is that someday I will finally
squeeze off an RFE to Franz: "Please, no full stops after identifiers in
error messages." The problem is AllegroCL error output appearing in the
listener, say:
Division by zero in: KT-COMPUTE-AVERAGE.
Normally I could click on the /output text/ and hit ctrl-alt-., but that
little full stop at the end of the error message, while a great comfort
i am sure to the former grammar teachers of Franz developers, prevents that.
Another thing is something like apropos being an interactive dialog in
which one can reshape the query and see diffferent result sets as I
type, not just the apropos output scrolling by in a listener.
Well, OK: also support. When CMUCL or SBCL stagger to their feets on
win32 will they come with installers? How about an automatic
"update-cmucl" function which reaches out to common-lisp.net to download
patches and apply them automatically? I love the idea of an entire
company (even if it is just one or two dedicated Lispniks) depending for
their survival on how slick are their tools.
>
> I'm particularly interested in IDE features that are applicable to an
> Emacs-like environment. I'm also curious to know what Lisp development
> environment the long-time lispers in this group use, and why.
First MCL, now AllegroCL. Just enough Lispworks and CMUCL to port things.
André Thieme
> I wish I'd never met Lisp. It has ruined me for all other
> programming languages. C and Java now look really ugly to me. XML
> is clearly someone's really bad joke at the world. Jelly is a new
> player in the Java/XML conspiracy. I'm not sure how to con someone
> into paying me to hack on Lisp.
Yes, it can be problematic to continue with something else.
For me personally the idea is to replace at least /some/ Java or C
coding with tools which translate from Lisp to human readable code in
Java or C. Perhaps I will try to do such a translator one day for PHP.
Then, another way could be to simply solve some tasks in Lisp. I have
some freedom, so if I do a solution for a tool in Lisp and make it
running and all this happens fast it would take some time until someone
finds out. And if someone found it out, he/she will probably not care
too much about it, as long it is doing a good job :)
> Lispworks costs $999 and MCL runs $750. I didn't look at Franz.
> I've had enough sticker shock for one day. I'm not buying a
> commercial Lisp soon, but I have downloaded the Lispworks Personal
> Edition to play with to see how I like that.
>
>
> I think I would like to have that integration. I would like to
> program in Lisp that way. I just don't have $999 to spend on it.
You can use the personal edition of LW for a lot of things. When you
feel ready you can buy the Pro version.
Someone (I think someone from one of the vendors) once said, that if
coding is your hobby you could buy a license. Some people whose hobby is
photography pay 5k for their equipment. Why not buying a commercial Lisp?
Maybe the same person criticised that too few people are willing to pay
for software more money than for the hardware, where the software will
run on.
Don't want to start a flame war.. but Kent Pitman seems to be right when
he is sceptical about free software (meaning the free beer part).
People get used to cheap or free software. This is not very good for our
buisiness. The sources can in many cases be open and in some fewer cases
the customer can also get a license which allows him to do his own
modifications. But anyway, the software should cost some money...
In LispWorks Pro for example we get the source code for the editor with
the right to do our modifications on it.
When I learned enough Lisp (in my subjective opinion) I will buy the Pro
version and sell some software. This should work for you too. If you
like Lisp and like to program it must be possible for you to spend some
time on projects which you sell, to get your 1k $.
André
--
Joe Marshall
> This is a rant. Please move on if you don't like rants. Move on
> especially quickly if you don't like rants that are ill thought out
> and clearly the ravings of a lunatic.
I'd better move on, because....
Hey, wait a minute, if I didn't like the insane ravings of lunatics,
why am I subscribed to comp.lang.lisp?
> I wish I'd never met Lisp. It has ruined me for all other
> programming languages. C and Java now look really ugly to me. XML
> is clearly someone's really bad joke at the world. Jelly is a new
> player in the Java/XML conspiracy.
Join the club. I find non-lisp systems nauseating.
> I'm not sure how to con someone into paying me to hack on Lisp.
It can be done. The past few years have been very nasty to all
programmers, but things are looking much better now. People who want
to hire Lisp hackers exist, but they often don't advertise (what's the
point?). When my company was hiring (back in 2000, before the bubble
burst) we *couldn't* find enough lisp hackers.
> Since I've gotten my Mac, I've been ruined for all other desktop
> systems. Sadly, there are a limited number of Lisp systems that work
> on it well.
At least they still make Macs. I miss my LispM.
> Lispworks costs $999 and MCL runs $750. I didn't look at Franz.
> I've had enough sticker shock for one day. I'm not buying a
> commercial Lisp soon, but I have downloaded the Lispworks Personal
> Edition to play with to see how I like that.
Franz is expensive, no question. But neither Franz nor Xanalys are
trying to gouge money out of hobbyists. The Franz and Xanalys `trial
editions' are pretty good and both Franz and Xanalys *encourage*
hobbyists to actively use them.
> Emacs + SLIME + SBCL take up less memory than Xcode. The startup
> time is shorter. The programming experience is far more
> interactive. Yes, it looks primitive by comparison. In reality, it
> seems a bit more sophisticated.
Yeah, looks are deceptive.
> The problem is, I am just not loving Emacs so much (mostly due to
> key chords). Yes, I use it for mail and gnews. That doesn't mean I
> like it.
Fix the chords.
> There is no Cocoa bridge and no .app packager. There is no
> integration to Xcode and its tools at all.
>
> I think I would like to have that integration. I would like to
> program in Lisp that way. I just don't have $999 to spend on it.
> Besides. I think that a worthy free Lisp implementation and dev
> environment for OS X is very useful for Lisp's future as a perhaps
> not so dead language. Obligatory Miracle Max quote goes here.
Hmmm, you want to hack lisp, you want to inteface Emacs + CommonLisp
to XCode. Did I mention that Emacs is written in Lisp? It isn't
Common Lisp, but it'll do in a pinch. I have an idea.
> So I've got a wish list. It's not a long list. Call it a short
> list.
>
> * SBCL should support threads on OS X using the same programmer level
> API as on Linux x86. If someone points me to where to look at
> hacking that in, I'd be happy to help.
>
> * SBCL should have a Cocoa bridge and Carbon wrappers for OS X.
> Again, I'm willing to work on this.
>
> * There should be a nicer IDE for Lisp that fits the OS X Aqua
> interface guidelines. Fewer key chords than Emacs has would be
> nice. SLIME like functionality is a must. The IDE should be part
> of the Lisp image. All forms of RPC suck. A threaded Lisp would
> help.
>
> * The Lisp IDE should be able to leverage Xcode for Cocoa, Carbon,
> Quicktime, and other documentation. App bundles and also
> frameworks should be things that you can make.
I have done no work whatsoever with SBCL, OS X, Cocoa, Carbon, Aqua,
SLIME, Xcode, or App bundles, so I cannot give you any help but
encouragement.
Dive in.
Just start.
Anywhere.
--
~jrm
André Thieme
>
> Apologies if this has been hashed out before, but what are the
> features that make the commercial Lisps so much better?
Support is a key. If someone develops nice libs it is highly probable
that they will make them work in LispWorks and Allegro.
I personally have not seen a lib that works on clisp and cmucl and
explicitely /not/ on the commercial ones.
If you want some modern feature it /could/ be in a free version while it
probably /is/ in the commercial versions.
One thing which someone else could confirm: the commercial versions
usually produce faster/smaller code.
André
--
Kenny Tilton
Edward Jason Riedy wrote:
> And Kenny Tilton writes:
> - Am I being obscure? Try this: if you cannot find $750 for a commercial
> - Lisp, you either are not really a programmer or you do not really like
> - Lisp. Your choice.
>
> Or you develop software that's expected to run on....
Sorry, my rant was aimed at one ranter's situation, tho I can see how it
could be taken as a blanket rant. Lispworks won't work for you?
In general, however, I /do/ get a kick out of the fact that CMUCL et al
exist, and in part because it can be a fallback for delivery should
licensing negotiations break down, but no self-respecting developer
tolerates unnecesary impediments to productivity. Hell, why do Lispniks
not mind (much) having to roll a set of bindings to get to some C
library? because Lisp makes them so much more productive the rest of the
time. Same idea with IDEs. If $1000 sounds like a lot, you just are not
writing that much code. No matter how much I loathe MS, the only thing
that matters to me is Lisp, and I have a great Lisp environment
available only on win32, so win32 it is. Until I snag a G5 just before
my developer discount expires, then it will be back to MCL (or
Lispworks) if Franz has not by then ported their IDE to Cello so I can
use their IDE on OS X.
Kenny Tilton
Tim Bradshaw wrote:
Sorry. I am being selfish. The Yobbos would be helping me with Cello if
they were not so busy with CMUCL and SBCL and OpenMCL. As it is, Barlow
said the earliest he could get involved would be this winter. Well, not
sure on that, i just remember something about "Hell freezing over".
Can't wait!
kenny
Espen Vestre
> Sorry, my rant was aimed at one ranter's situation, tho I can see how
> it could be taken as a blanket rant. Lispworks won't work for you?
I repeat (and this time I checked!): Runtimes for LW for UNIX are
still charged for! (list price is now USD 900 in the US)
--
(espen)
Joe Marshall
> Kenny Tilton <kti...@nyc.rr.com> writes:
>
>> If the light does not go on, I suggest you arrange for another
>> twenty-thousand man-hours priced at ten cents each so you can get
>> CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial
>> lisps.
>
> Apologies if this has been hashed out before, but what are the
> features that make the commercial Lisps so much better?
The biggest feature is stability. Lispworks and Allegro each are
platforms that work fairly uniformly across their various ports. At
Content Integrity we did the bulk of our development on Windows, but
deployed the product on HP-UX (what the customer wanted). There were
some portability issues, but they were reasonably minor and isolated
to the low-level pathname hacking and database API.
The second is support. When you are running a company, you have to
make tradeoffs between doing things yourself or getting someone else
to do it. With a commercial lisp, the only barrier is money. With
sufficient cash, you can get almost the entire Franz corporation
working on solving your problem. And they are all a phone call away.
The third is expertise. The people who work for these companies don't
need a day-job to pay the bills, so they can spend 40+ hours a week on
their product. There simply aren't as many available man-hours to put
in to a non-commercial lisp.
> I'm particularly interested in IDE features that are applicable to an
> Emacs-like environment. I'm also curious to know what Lisp development
> environment the long-time lispers in this group use, and why. For
> instance, Joe Marshall, how do you hack your Lisp nowadays?
XEmacs 21.4.13 on Windows 2K at home, on WinXP at Northeastern.
The reason is that XEmacs not only does lisp, but it does every other
language I have the misfortune to use (C, C++, C#, assembly, Java,
HTML, XML, etc.) *and* handles my mail and news *and* acts as terminal
for the shell *and* has TAGS and grep *and* is customizable.
The graph tools in Lispworks can be quite useful for inspecting
objects and classes, so I'll occasionally run them, but I rarely edit
code in anything but XEmacs.
--
~jrm
Christian Lynbech
Espen> I repeat (and this time I checked!): Runtimes for LW for UNIX are
Espen> still charged for! (list price is now USD 900 in the US)
Does the "UNIX" part include Linux and/or Mac OSX?
Kaz Kylheku
> I'm not sure how to con someone into paying me to hack on Lisp.
CONS omeone? Hmm... ;)
Friedrich Dominicus
>>>>>> "Espen" == Espen Vestre <espen@*do-not-spam-me*.vestre.net> writes:
>
> Espen> I repeat (and this time I checked!): Runtimes for LW for UNIX are
> Espen> still charged for! (list price is now USD 900 in the US)
>
> Does the "UNIX" part include Linux and/or Mac OSX?
No.
Regards
Friedrich
--
Please remove just-for-news- to reply via e-mail.
Dave Fayram
>
> > I wish I'd never met Lisp. It has ruined me for all other
> > programming languages.
>
>
> > Lispworks costs $999 and MCL runs $750.
>
> months to afford a $1600 Apple frickin II in 1978 so I could do integer
> basic in a whopping 16K of RAM, and I was happy to have it.
>
> If the light does not go on, I suggest you arrange for another
> twenty-thousand man-hours priced at ten cents each so you can get
> CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial lisps.
>
> <sigh> C'mon, Lispniks are supposed to be /intelligent/.
>
> Am I being obscure? Try this: if you cannot find $750 for a commercial
> Lisp, you either are not really a programmer or you do not really like
> Lisp. Your choice.
Kenny, I very much like Lisp, and I am very much a programmer.
However, unlike most folks here who have that exalted "senior" status,
I am fresh out of college. I sat around for 9 months begging for a
job, moving from couch to couch, accumulating debt and literally doing
website design and maintenance so that I could pay for gas to get to
my next crappy job interview.
Finally Lockheed Martin comes along and says, "Well hire you at 15%
less than the market rate, but we promise you two raises per year for
the first two years, and a big bonus at the end of the first year." I
frantically grabbed on because I LIKE EATING AND OWNING A RESIDENCE.
I am still trying to climb out of the hole I sunk into. Now, if you're
like me, you always have tons of good ideas, possibly even
money-making ones. When I found Lisp, I thought, "Gee, this would be
perfect for (idea here)." Too bad the commercial lisps (which would
bring platform mobility, a key feature for many of my ideas) are so
far out of my pathetic price range that I can't even dream of them.
Maybe I could have used the bonus or raise money? Bzzt, LMCO effect, I
don't get them now due to "Policy Changes". Apparently my contract has
a nice clause allowing that. Too bad I was too delirious from lack of
food to really understand it when I signed it.
Yes, Kenny, I'm sure David recognizes the economic value of the
commercial lisps. I do too. I am still totally unable to purchase
them. I'm already living on rice to make ends meet. There is no more
fat to trim.
I too, am sorry I met lisp, because it may be years before I can
really get into it, and right now it's all I want to do. I still enjoy
it, but lately it seems like the better parts of lisp are a world only
rich people or companies can enjoy.
- dlf
Kenny Tilton
OK, sorry, I missed that. I still would develop with the most productive
tool and then deliver on something free if negotiations with Xanalys
were unsatisfactory. Of course I would be using Cello, not CAPI, so I
would have that option.
:)
Wade Humeniuk
> And Kenny Tilton writes:
> - Am I being obscure? Try this: if you cannot find $750 for a commercial
> - Lisp, you either are not really a programmer or you do not really like
> - Lisp. Your choice.
>
> processors and cannot afford to force your clients / users to
> afford $750 per processor.
>
Its a big machine. How many millions did it cost?
http://hpcf.nersc.gov/computers/SP/
Besides that how much does it cost to maintain the sucker?
Power consumption and cooling probably cost a few million
a year. Then what does it cost for the programmers, sysadmins,
and other management overhead? Anyone that needs a machine
that expensive has to pay to keep that machine running. Its
a reality that to use that machine you have to pay.
For a million dollars, maybe even much less you could get a
Lisp running on it. (Do you need a Lisp that compiles an app
for parallel execution?) Why does the Lisp have to be free?
Wade
Wade Humeniuk
>
>
> Kenny, I very much like Lisp, and I am very much a programmer.
> However, unlike most folks here who have that exalted "senior" status,
> I am fresh out of college. I sat around for 9 months begging for a
> job, moving from couch to couch, accumulating debt and literally doing
> website design and maintenance so that I could pay for gas to get to
> my next crappy job interview.
>
Are you too poor to own a computer, or have an internet connection?
Did you pay for it? It is no different. At least you can get the
LispWorks Personal Edition for free and do useful work. Harlequin
has been very generous. Or get one of the free Lisps, there is
no need to play the victim.
Wade
Kenny Tilton
Dave Fayram wrote:
> Kenny Tilton <kti...@nyc.rr.com> wrote in message news:> David Steuber wrote:
>
>>>I wish I'd never met Lisp. It has ruined me for all other
>>>programming languages.
>>
>>...and...
>>
>>
>>>Lispworks costs $999 and MCL runs $750.
>>
>>Sit in the lotus position until the light goes on. Hint: I starved for
>>months to afford a $1600 Apple frickin II in 1978 so I could do integer
>>basic in a whopping 16K of RAM, and I was happy to have it.
>>
>>If the light does not go on, I suggest you arrange for another
>>twenty-thousand man-hours priced at ten cents each so you can get
>>CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial lisps.
>>
>>It is /so/ important that our tools be free, no matter what the cost.
>><sigh> C'mon, Lispniks are supposed to be /intelligent/.
>>
>>Am I being obscure? Try this: if you cannot find $750 for a commercial
>>Lisp, you either are not really a programmer or you do not really like
>>Lisp. Your choice.
>
>
> Kenny, I very much like Lisp, and I am very much a programmer.
> However, unlike most folks here who have that exalted "senior" status,
> I am fresh out of college. I sat around for 9 months begging for a
> job, moving from couch to couch, accumulating debt and literally doing
> website design and maintenance so that I could pay for gas to get to
> my next crappy job interview.
>
> Finally Lockheed Martin comes along and says, "Well hire you at 15%
Congratulations!
I do not think we disagree at all. If while at Lockheed you come up with
a great idea for a home hobby project, you will probably use Lisp. If it
is going to mean thousands of hours of work, and if you have played with
trial versions of commercial tools, when you start using free toys you
will probably decide they are not so free. Unless you are a Linux fan,
in which case you likely no longer realize how much time you $pend on
free (as in "my time is worthle$$") software. But if you like using an
installer for three minutes instead of GCC for three days, you will
likely find a way to come up with the bucks.
But right now you do not have any Lisp code to write, and that is my
point. Now $1000 or so looks crazy[1]. Fortunately you have trial
versions which cost nothing, are barely crippled, and come with installers!
Good luck with the gig.
kenny
[1] Unless you realize that Lisp jobs are probably just a year away and
you want to have more than c.l.lisp on your resume when the hiring
starts. :)
Jock Cooper
> David Steuber <da...@david-steuber.com> writes:
>
> > player in the Java/XML conspiracy. I'm not sure how to con someone
> > into paying me to hack on Lisp.
>
> You may start asking around among your friends, relatives and
> colleagues who are not computer savvy. Some of them probably need a
> small application or tool. And, provided it works, they probably
> don't care which language it's written in.
>
> In short, start with grocery store size businesses. You may not
> become rich quick, but this may bootstrap a for pay Lisp coding
> activity.
>
All the C.L.L. denizens who are wishing for CL jobs should get
together and start a company (think: telecommute). What would happen
if all this productivity was focused...
André Thieme
Sometimes I ask myself: "How it comes that at the end of money still so
much month is left?"
André
--
André Thieme
Okay, I am the boss!
André
--
Tim Bradshaw
> What, you mean like these?
>
> <URL:http://www.helispot.com/photos/03269.html>
> <URL:http://www.bellhelicopter.textron.com/en/aircraft/military/
> bellAH-1Z.cfm>
> <URL:http://www.rotaryaction.com/images/clearapd.jpg>
>
Well sort of. Ours are invisible, of course, and come armed with death rays.
Peter Herth
> Besides that how much does it cost to maintain the sucker?
> Power consumption and cooling probably cost a few million
> a year. Then what does it cost for the programmers, sysadmins,
> and other management overhead? Anyone that needs a machine
> that expensive has to pay to keep that machine running. Its
> a reality that to use that machine you have to pay.
>
> For a million dollars, maybe even much less you could get a
> Lisp running on it. (Do you need a Lisp that compiles an app
> for parallel execution?) Why does the Lisp have to be free?
Of course it seems to be ridiculous to complain about the
costs of Lisp if you look at that machine - however, even
with big budgets for a project, however big they are, if
you cannot convince the one who does the financial decisions
that the money spent for a Lisp is a good idea, you are
out of luck :( We need a lot of Lisp success stories until
we can hope to spend money on Lisp at all. (At least for
certain kinds of projects)
Peter
--
Peter Herth
pet project: http://dawn.netcologne.de
homepage: http://www.peter-herth.de
lisp stuff: http://www.peter-herth.de/lisp.html
get Ltk here: http://www.peter-herth.de/ltk/
Gisle Sælensminde
>
>
> Espen Vestre wrote:
>> Kenny Tilton <kti...@nyc.rr.com> writes:
>>
>>
>>>Sorry, my rant was aimed at one ranter's situation, tho I can see how
>>>it could be taken as a blanket rant. Lispworks won't work for you?
>>
>>
>> I repeat (and this time I checked!): Runtimes for LW for UNIX are
>> still charged for! (list price is now USD 900 in the US)
>
> OK, sorry, I missed that. I still would develop with the most productive
> tool and then deliver on something free if negotiations with Xanalys
> were unsatisfactory. Of course I would be using Cello, not CAPI, so I
> would have that option.
Actually the price is $5500 for a commecial unix license ($3300 for an
academic license), according to the PDF on their web page, which is quite
expensive. The price at least made me think twice before buying.
Since the solaris servers will be replaced next year, I did not get further
to get it purchased. The price of $550 for the Linux academic licence would
not have beens such an issue. Not that I blame them, there is probably not
that many solaris sales.
--
Gisle Sælensminde
Computational biology unit, University of Bergen, Norway
Email: gi...@cbu.uib.no
Gorbag
"Erann Gat" <gNOS...@flownet.com> wrote in message
news:gNOSPAMat-C0A7F...@nntp1.jpl.nasa.gov...
> In article <lhacz2p...@dodo.bluetail.com>,
> > > twenty-thousand man-hours priced at ten cents each so you can get
> > > CMUCL/SBCL+ILISP/SLIME working nowhere near as well as commercial
> > > lisps.
> >
> > features that make the commercial Lisps so much better?
>
This is a key insight, and one of the reasons I stuck to the LispMs so long
(for ZMACS).
>
> It's like EMACS, but it's written in (and more to the point, hackable
> in) Common Lisp. So I don't have to learn a new dialect of Lisp to
> program my editor.
Plus, the editor and your app are running IN THE SAME ENVIRONMENT. So they
can influence each other - your editor can know what's in your environment,
directly grok the process list, etc. ILisp, SLIME, etc. all try to do this
through a port but never really get to full integration the way FRED and
ZWEI do.
>
> Fred is the one feature that has kept me a loyal MCL user for nearly
> twenty years. I started with the first release of Coral Common Lisp and
> I've never looked back.
>
> E.
Paolo Amoroso
> I am still trying to climb out of the hole I sunk into. Now, if you're
> like me, you always have tons of good ideas, possibly even
> money-making ones. When I found Lisp, I thought, "Gee, this would be
> perfect for (idea here)." Too bad the commercial lisps (which would
> bring platform mobility, a key feature for many of my ideas) are so
> far out of my pathetic price range that I can't even dream of them.
For the record, Paul Graham based a business worth over 50 M$ on CLISP
and other free tools (e.g. Perl, if I recall correctly). Graham chose
CLISP over Allegro CL because, at the time, the performance of CLISP
was better than Allegro CL's for the kind of application Graham was
interested in.
By the way, this motivated Franz to improve their product. The simple
streams feature was designed to address the problem.
Paolo
--
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Recommended Common Lisp libraries/tools (Google for info on each):
- ASDF/ASDF-INSTALL: system building/installation
- CL-PPCRE: regular expressions
- UFFI: Foreign Function Interface
David Steuber
> David Steuber wrote:
>
> >
> > Since I've gotten my Mac, I've been ruined for all other desktop
> > systems.
>
Enough for two LispWorks professionl licenses ;-)
It probably wan't exactly a wise expense either.
If I was confident that I had a good idea for a commercial app, I
would spring for LispWorks or MCL in a second. Well, if I had the
cash on hand I would. I spent it all on the Mac a bit over a year ago
now.
I'm not so much ranting against Lisp as ranting against Lisp not being
the mainstream language of choice for everyone. If Lisp was
mainstream, Xcode would do it. I expect Dev Studio would do it also.
Somewhere, something went terribly wrong. I still think that if a
good, free Lisp... scratch that... a better free Lisp was available
Lisp would be a healthier language. I can certainly work through
examples in OnLisp, ACL, PAIP, and a few others using the free Lisps.
Also, SLIME does help a fair bit. SBCL, OpenMCL, CLisp, SLIME, et al
are moving along. They aren't standing still. They are useful. But
they do have limits.
Anyway, I didn't intend to start a Free vs Commercial debate.
Commercial companies do what they do to make money and that is as it
should be. To be fair, I have not accumulated loads of experience
using GCC to compile C code written using Emacs or Vim. Free C is no
better off other than people are inexplicably using C for applications
programs. Well, that's not quite true. There are free IDEs for
C/C++.
If that isn't bad enough, my local news feed has been screwy today and
I am relying on Google to see followups to my post.
David Steuber
>
Well, I think I might be able to help you with your math error.
David Steuber
> You can use the personal edition of LW for a lot of things. When you
> feel ready you can buy the Pro version.
> Someone (I think someone from one of the vendors) once said, that if
> coding is your hobby you could buy a license. Some people whose hobby is
> photography pay 5k for their equipment. Why not buying a commercial Lisp?
> Maybe the same person criticised that too few people are willing to pay
> for software more money than for the hardware, where the software will
> run on.
I've already downloaded the LW personal edition. If it impresses me
enough, I'll start puting change in a jar to buy the pro version. I'm
not against paying for software (otherwise I wouldn't have a Mac).
Even as a hobby, software can be worth paying money for. I just don't
have it right now.
I might well be able to enjoy programming Lisp as a hobby without any
of the commercial ware. The advantage I see in the commercial ware is
that it makes it easier to share (or sell) my gems.
> Don't want to start a flame war.. but Kent Pitman seems to be right when
> he is sceptical about free software (meaning the free beer part).
> People get used to cheap or free software. This is not very good for our
> buisiness. The sources can in many cases be open and in some fewer cases
> the customer can also get a license which allows him to do his own
> modifications. But anyway, the software should cost some money...
> In LispWorks Pro for example we get the source code for the editor with
> the right to do our modifications on it.
I expect market forces will take care of all that. It has been
pointed out many times that free software carries a heavy time cost.
However, as good as Lisp is it still has to compete against other
languages. There seems to be no shortage of people who are satisfied
with lower quality products.
> When I learned enough Lisp (in my subjective opinion) I will buy the Pro
> version and sell some software. This should work for you too. If you
> like Lisp and like to program it must be possible for you to spend some
> time on projects which you sell, to get your 1k $.
Well, never say never. I'm sure I would be very happy if I could get
$1k in revenue from selling software of my own. I would be happier
still if I could live off the proceedes.
Whether my tools are free or I pay a license fee is not really the
point for me when it comes to distributing code. I can distribute raw
source with the "you build it if you can" attitude. If I am going to
distribute binaries I expect them to work on systems other than the
machine I compiled on.
Whatever the economics work out to be in the long run, I expect that
keeping down the barriers to entry can only help the Lisp community as
a whole. Presumably people who have a positive experience with the
language will want to use it.
Xavier Maillard
[ ... ]
> Don't want to start a flame war.. but Kent Pitman seems to be
> right when he is sceptical about free software (meaning the
> free beer part). People get used to cheap or free software.
> This is not very good for our buisiness. The sources can in
> many cases be open and in some fewer cases the customer can
> also get a license which allows him to do his own
> modifications. But anyway, the software should cost some
^^^^^^
I totally don't agree with that. A software _could_ cost but
this is non sense to argue all software should cost something.
On the other hand I totally agree with you wehn you say people
are getting used not to pay for software in our Free Software
world. But as Stallman and other say, Free Software doesn't mean
it should be free (no charge). It is common that Free Software
doesn't cost anything but it is much a matter of choice than
anything else.
As a Hacker (in the good sense of the term), hacking free
software is much more a hobby than something else. I love
developping and sharing with other and nevermind to be paid for
that hobby.
Here in France, we have several company doing Free Software and
the majority of them do it for free. What they make thei
customers pay for is the service: support, installation,
maitainance, ...
Regards,
P.S: sorry for the bad english expression ;)
--
Xavier Maillard| "Stand Back! I'm a programmer!"
.0. ze...@gnu-rox.orgz|
..0 (+33) 326 770 221 | Webmaster, emacsfr.org
000 PGP : 0x1E028EA5 | Membre de l' APRIL
Joe Marshall
> I'm not so much ranting against Lisp as ranting against Lisp not being
> the mainstream language of choice for everyone. If Lisp was
> mainstream, Xcode would do it. I expect Dev Studio would do it also.
It's only waiting for someone to make the appropriate COM interfaces.
Julian Stecklina
> Christian Lynbech <christia...@ericsson.com> writes:
>
>>>>>>> "Espen" == Espen Vestre <espen@*do-not-spam-me*.vestre.net> writes:
>>
>> Espen> I repeat (and this time I checked!): Runtimes for LW for UNIX are
>> Espen> still charged for! (list price is now USD 900 in the US)
>>
>> Does the "UNIX" part include Linux and/or Mac OSX?
> No.
What about FreeBSD?
Regards,
--
Julian Stecklina
Signed and encrypted mail welcome.
Key-Server: pgp.mit.edu Key-ID: 0xD65B2AB5
FA38 DCD3 00EC 97B8 6DD8 D7CC 35D8 8D0E D65B 2AB5
Any sufficiently complicated C or Fortran program
contains an ad hoc informally-specified bug-ridden
slow implementation of half of Common Lisp.
- Greenspun's Tenth Rule of Programming
Lupo LeBoucher
Wade Humeniuk <whumeniu-delete-th...@telus.net> wrote:
>Edward Jason Riedy wrote:
>> And Kenny Tilton writes:
>> - Am I being obscure? Try this: if you cannot find $750 for a commercial
>> - Lisp, you either are not really a programmer or you do not really like
>> - Lisp. Your choice.
>>
>> Or you develop software that's expected to run on 100-9000
>> processors and cannot afford to force your clients / users to
>> afford $750 per processor.
>>
>
>Its a big machine. How many millions did it cost?
>
>http://hpcf.nersc.gov/computers/SP/
>
>Besides that how much does it cost to maintain the sucker?
>Power consumption and cooling probably cost a few million
>a year.
You labor under the delusion that NERSC has money flowing out of its
bung-hole to spend making a couple of Lisp enthusiasts happy, even if it
does make their life easier. Grad students don't cost anything, and you
can get them to write the code in whatever you like.
It's government science, fer chrissakes.
While the military might have such money, they ain't gonna put it on their
Crays or IBM SP's either, unless you can justify why you're not writing it
in C++ or Fortran. If you can't do that at whatever YoYo-Dyne company you
work for, imagine how much harder it is to pimp it to the GI Joes, or
pointy-headed underfunded numerics guys who, if they know anything about
Lisp at all, remember their Lisp machines garbage collecting and thrashing
like a sickly washing machine all day long.
But yeah, it would be nice if you could rent Lisp time on that kind of
heavy metal; especially if you don't have to use a GPLed 1980s IDE to
make it do tricks.
-Lupo
"there is often a certain nostalgia among us moderns for the memory of men
and women who could, for a time at least, shoulder fate aside and worship
Thor, because they believed in their own strength" -E.J. Sharpe <i...@io.com>
Julian Stecklina
> Plus, the editor and your app are running IN THE SAME ENVIRONMENT. So they
> can influence each other - your editor can know what's in your environment,
> directly grok the process list, etc. ILisp, SLIME, etc. all try to do this
> through a port but never really get to full integration the way FRED and
> ZWEI do.
Just for the record: ACLs internal editor is crap. ACL+xeli+xemacs
nice for editing, but ACL+SLIME is *very* comfortable. One thing
that's missing is, that you cannot use the ACL GUI to navigate to a
particular source code location.
Tyler Eaves
I agree with much of that. One suggestion, grab a free app called
uControl, which allows you to remap the keyboard. The particular
trick you want is to remap your caps lock key to be a Control key.
This makes Emacs far more pleasant to use, as you now have a control
key practically under your left pinky, as all good keyboards should.
Edward Jason Riedy
- While the military might have such money, they ain't gonna put it on their
- Crays or IBM SP's either, unless you can justify why you're not writing it
- in C++ or Fortran.
Just as important in my case is justifying why you cannot just
link the library into existing C++ and Fortran codes without
buying a single, specific CL implementation and restructuring
your code around it. Allegro has an appropriate C->Lisp
ability, but it's not portable between CL implementations (I'd
love to be wrong), and I haven't timed it for working on large
data sets. Plus, you have to link the C++/C/Fortran routine
into the Lisp image, which won't work for users who already
have execution frameworks.
I'd also have to explain why you should forget about running
on an SV1 (Cray vector machine) or Altix (SGI IA64 DSM box).
Scientific HPC isn't a market for companies that want to make
money, so I'm not expecting a commercial Lisp on current or
future platforms of interest.
(And no amount of Lispworks niceness matters. Allegro's the
one with the right support for some of the platforms I can use,
and they have a free beta of their next version. I have until
Oct. to find a "killer app" use.)
- [...], or pointy-headed underfunded numerics guys who, if they know
- anything about Lisp at all, [...]
Well, some of the numerics folks around here have good enough
knowledge about Lisp. ;)
Sorry I'm ranging so far afield; just feel the need to semi-rant.
Jason
--
David Steuber
I've got it (1.4.4). It does improve things tremendously. Using the
touchpad as a scroll wheel is also useful from time to time. I did
have a minor hitch when I forgot to remove 1.4.3 before upgrading from
10.3.3 to 10.3.4. Things seem to be working fine now after a wee bit
of fiddling around.
I've got LW personal edition installed, but I haven't tried it out
yet. Been busy today. I'm also waiting on e-mail from Franz to grab
Allegro 6.2 Trial Edition.
It's too bad the caps lock light isn't affected by uControl. I hate
seeing it on. I guess even a kernel module doesn't have access to
that though.
I still haven't tackled CLOS yet, but I will be very interested to see
how one goes about subclassing Objective-C classes. More interesting
would be multiple inheritance. One thing I haven't quite grokked yet
with Objective-C, or at least with Cocoa, is when you are supposed to
release an object so that it will get deallocated properly. I'm not
quite clear on the protocol. I should probably do some Objective-C
programming just to get sorted out on that point, but like I said with
the first post, Common Lisp has ruined me. Objective-C isn't any
prettier than C. It has even more nasty syntax stuff. Bleh!
Actually, a pet peeve I have about Java is the fact that you call new
to create an object but there is no delete to balance it like with
C++. If you have automatic memory management, why do you need new?
It should be implicit. I suppose one might have the same complaint
about make-array or make-hash-table, but somehow I don't see anything
wrong with those functions.
Common Lisp seems to have anticipated most needs. Those libraries
people complain about not being present (myself included) for
threading, sockets, guis, etc, are missing from the free Lisps. Well,
only if you ask for all the features at once. SBCL has sb-bsd-socket.
sb-thread only works on x86 Linux at the moment, but OpenMCL has
posix threading on OS X. There is a project called McCLIM for GUI.
Obviously it is tricky to do a cross platform GUI. One could argue
that Java's attempt was not wholy successful. Also, last I looked,
there was a bit of a schism between Swing and AWT.
The GUI doesn't matter so much if you are doing web services. If you
have decided that OS X rocks and you don't even want to look at
Windows again (and maybe not even X11), then the GUI doesn't need to
be portable. It's probably easier to use the MVC pattern and keep the
GUI as seperate as possible anyway if you want to port to something
using a slightly different event model.
If I never met Lisp, I would probably be cursing Objective-C and Java
as I coded in them with no hope of something better.
Friedrich Dominicus
>>>
>>> Does the "UNIX" part include Linux and/or Mac OSX?
>> No.
>
> What about FreeBSD?
AFAIK does there exist not "native" LispWorks for FreeBSD. So if one
run LispWorks one will probably run the Linux version.
Brian Downing
David Steuber <da...@david-steuber.com> wrote:
> I've got it (1.4.4). It does improve things tremendously. Using the
> touchpad as a scroll wheel is also useful from time to time. I did
> have a minor hitch when I forgot to remove 1.4.3 before upgrading from
> 10.3.3 to 10.3.4. Things seem to be working fine now after a wee bit
> of fiddling around.
For a scroll wheel a far more elegent solution (instead of the uControl
scroll wheel emulation) is to get SideTrack
(http://www.ragingmenace.com/software/sidetrack/), which is a new driver
that enables some of the advanced features of the trackpad hardware that
Apple uses. This includes horizontal and vertical scroll areas (on any
side), and other buttons and functions tied to corner taps.
I have mine set up to generate scroll wheel events if I track along the
right side of the pad, and have mouse2, mouse3, Expose all windows, and
Expose desktop set up on taps to the four corners.
Unfortunatly, the current version (0.8 beta) expires in 11 days, and
there's been no new release! Hopefully one gets made before it expires,
because I'd hate to go back to the Apple trackpad driver.
-bcd
--
*** Brian Downing <bdowning at lavos dot net>
Peter Herth
> Actually, a pet peeve I have about Java is the fact that you call new
> to create an object but there is no delete to balance it like with
> C++. If you have automatic memory management, why do you need new?
> It should be implicit. I suppose one might have the same complaint
> about make-array or make-hash-table, but somehow I don't see anything
> wrong with those functions.
Well, with an implicit new, any List structure would be in infinite
size if you have:
class Cons
{
Object car;
Cons cdr; // <= if this allocated a Cons object by default, this would mean
trouble :)
}
> Common Lisp seems to have anticipated most needs. Those libraries
> people complain about not being present (myself included) for
> threading, sockets, guis, etc, are missing from the free Lisps. Well,
> only if you ask for all the features at once. SBCL has sb-bsd-socket.
> sb-thread only works on x86 Linux at the moment, but OpenMCL has
> posix threading on OS X. There is a project called McCLIM for GUI.
> Obviously it is tricky to do a cross platform GUI. One could argue
> that Java's attempt was not wholy successful. Also, last I looked,
> there was a bit of a schism between Swing and AWT.
It may be tricky do to a cross platform GUI, but is has been done! :)
Look at Ltk: http://www.peter-herth.de/ltk/ As of Version 0.8.3
openMCL is amongst the supported Lisps and it runs nicely under
a lot of other Lisps and operating systems. A peek how it looks
under OS X: http://www.peter-herth.de/ltk/ltk-osx.png
Peter Herth
> Actually, a pet peeve I have about Java is the fact that you call new
> to create an object but there is no delete to balance it like with
> C++. If you have automatic memory management, why do you need new?
> It should be implicit. I suppose one might have the same complaint
> about make-array or make-hash-table, but somehow I don't see anything
> wrong with those functions.
Well, with an implicit new, any List structure would be in infinite
size if you have:
class Cons
{
Object car;
Cons cdr; // <= if this allocated a Cons object by default, this would mean
trouble :)
}
Or did I misunderstood you ?
> Common Lisp seems to have anticipated most needs. Those libraries
> people complain about not being present (myself included) for
> threading, sockets, guis, etc, are missing from the free Lisps. Well,
> only if you ask for all the features at once. SBCL has sb-bsd-socket.
> sb-thread only works on x86 Linux at the moment, but OpenMCL has
> posix threading on OS X. There is a project called McCLIM for GUI.
> Obviously it is tricky to do a cross platform GUI. One could argue
> that Java's attempt was not wholy successful. Also, last I looked,
> there was a bit of a schism between Swing and AWT.
It may be tricky do to a cross platform GUI, but is has been done! :)
Look at Ltk: http://www.peter-herth.de/ltk/ As of Version 0.8.3
openMCL is amongst the supported Lisps and it runs nicely under
a lot of other Lisps and operating systems. A peek how it looks
under OS X: http://www.peter-herth.de/ltk/ltk-osx.png
Peter
--
Rob Warnock
+---------------
| I'd also have to explain why you should forget about running
| on an SV1 (Cray vector machine) or Altix (SGI IA64 DSM box).
What's wrong with the SGI Altix? It can run in cc/NUMA mode
(completely cache-coherent across all processors & memories)
as well as in partitioned mode (several smaller cc/NUMA images
non-coherently DSM'd together).
-Rob
-----
Rob Warnock <rp...@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
Paolo Amoroso
> While the military might have such money, they ain't gonna put it on their
> Crays or IBM SP's either, unless you can justify why you're not writing it
> in C++ or Fortran. If you can't do that at whatever YoYo-Dyne company you
> work for, imagine how much harder it is to pimp it to the GI Joes, or
> pointy-headed underfunded numerics guys who, if they know anything about
> Lisp at all, remember their Lisp machines garbage collecting and thrashing
> like a sickly washing machine all day long.
Back in the Desert Storm days, LispMs garbage collecting and trashing
didn't look to the military like a major issue.
Paolo Amoroso
> Just for the record: ACLs internal editor is crap. ACL+xeli+xemacs
> nice for editing, but ACL+SLIME is *very* comfortable. One thing
Which, together with the mention of Franz engineer Charley Cox in
another thread, reminds me of something funny I noticed at a Franz
seminar a few months ago.
Charley did a demo of a web services application based on the
Webaction framework. The demo involved a few rounds of interactive
changes to the code.
He used a Windows laptop, and started showing the Allegro CL IDE.
Then switched to Emacs for the rest of the demo :)
Edward Jason Riedy
- What's wrong with the SGI Altix?
I wasn't maligning the Altix line; they're damned impressive.
I was noting that I'm not aware of an optimized, compiled Lisp
for it. Am I wrong? I'd love to be wrong... (I don't have
access to an Altix at the moment, but it's a target.)
Jason
--
Kenny Tilton
Paolo Amoroso wrote:
> Julian Stecklina <der_j...@web.de> writes:
>
>
>>Just for the record: ACLs internal editor is crap. ACL+xeli+xemacs
>>nice for editing, but ACL+SLIME is *very* comfortable. One thing
>
>
> Which, together with the mention of Franz engineer Charley Cox in
> another thread, reminds me of something funny I noticed at a Franz
> seminar a few months ago.
>
> Charley did a demo of a web services application based on the
> Webaction framework. The demo involved a few rounds of interactive
> changes to the code.
>
> He used a Windows laptop, and started showing the Allegro CL IDE.
> Then switched to Emacs for the rest of the demo :)
If you want to broaden the sample, there may be a videotape of me during
a demo trying to write and run a five-line function using
Emacs+ILisp+CMUCL with an audience of cmuclites and emaxen shouting out
key combos for my consideration. :)
otoh, when I use the Allegro IDE (on win32) in public, people just want
to drag me outside and burn me at the stake for practicing witchcraft.
tedious that. all demo long it's show them a feature, fight back the
mob, show then a feature, fight back the mob...
:)
kenny
--
Home? http://tilton-technology.com
Cells? http://www.common-lisp.net/project/cells/
Cello? http://www.common-lisp.net/project/cello/
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Your Project Here! http://alu.cliki.net/Industry%20Application
Paul F. Dietz
> I wasn't maligning the Altix line; they're damned impressive.
> I was noting that I'm not aware of an optimized, compiled Lisp
> for it. Am I wrong? I'd love to be wrong... (I don't have
> access to an Altix at the moment, but it's a target.)
Is that one of the debian targets? gcl is compiled and runs on all
the debian target architectures.
Paul
Rob Warnock
+---------------
Hmmm... According to <URL:http://www.sgi.com/servers/altix/configs.html>
they only support "SUSE LINUX Enterprise Server 8 + Service Pack 3", and
that only with "SGI ProPack" added. Does GCL run on SuSE? In 64-bit mode?
Also, does GCL support multiprocessors?
Edward Jason Riedy
- Does GCL run on SuSE? In 64-bit mode?
I don't believe it uses anything Debian-specific. And how
could anything operate in "32-bit mode" on an IA64? It's
native, not x86. I don't know if it supports >2GB arrays,
but I can live with that for now.
- Also, does GCL support multiprocessors?
There's an MPI interface, which suits my needs. wheee!!!
Now if I can port it to recent AIX with its slightly strange
mix of object formats, well, that's another question. As far
as the vector boxes go, at least I can do fun development to
get the software architecture straight.
Thank you again to the previous poster!
Jason
--
David Steuber
> David Steuber wrote:
>
> > Actually, a pet peeve I have about Java is the fact that you call new
> > to create an object but there is no delete to balance it like with
> > C++. If you have automatic memory management, why do you need new?
> > It should be implicit. I suppose one might have the same complaint
> > about make-array or make-hash-table, but somehow I don't see anything
> > wrong with those functions.
>
> Well, with an implicit new, any List structure would be in infinite
> size if you have:
>
> class Cons
> {
> Object car;
> Cons cdr; // <= if this allocated a Cons object by default, this would mean
> trouble :)
> }
>
> Or did I misunderstood you ?
Not entirely. The example above may or may not lead to trouble. The
cdr field could have allocation for it delayed until something is
actually assigned to it. That's an implementation issue that may or
may not be solvable in the general case. But I suppose if you can
solve it for some trivial case that all other cases build upon, then
you wouldn't need new.
> > Common Lisp seems to have anticipated most needs. Those libraries
> > people complain about not being present (myself included) for
> > threading, sockets, guis, etc, are missing from the free Lisps. Well,
> > only if you ask for all the features at once. SBCL has sb-bsd-socket.
> > sb-thread only works on x86 Linux at the moment, but OpenMCL has
> > posix threading on OS X. There is a project called McCLIM for GUI.
> > Obviously it is tricky to do a cross platform GUI. One could argue
> > that Java's attempt was not wholy successful. Also, last I looked,
> > there was a bit of a schism between Swing and AWT.
>
> It may be tricky do to a cross platform GUI, but is has been done! :)
> Look at Ltk: http://www.peter-herth.de/ltk/ As of Version 0.8.3
> openMCL is amongst the supported Lisps and it runs nicely under
> a lot of other Lisps and operating systems. A peek how it looks
> under OS X: http://www.peter-herth.de/ltk/ltk-osx.png
I'll have a look at it. I'm sure Ltk is great for X11 apps where
there is no notion of a CUA. On OS X though, does Tk handle the Aqua
guidlines appropriately? And is X11 required? That is really an
optional install on OS X and has only been bundled in since 10.3.
I didn't start using OS X until 10.2 and I had to manually fetch a
beta version of X11.app. I was happy to see it included with Panther
(even as an optional install). However, all X11 apps look like a
single instance of X11 to an OS X user.
These things won't stop me from playing with Ltk. But if I were to
write an OS X application that I intended for distribution to the
general population, I would want to use Aqua either via Carbon or
Cocoa.
Peter Herth
> I'll have a look at it. I'm sure Ltk is great for X11 apps where
> there is no notion of a CUA. On OS X though, does Tk handle the Aqua
> guidlines appropriately? And is X11 required? That is really an
> optional install on OS X and has only been bundled in since 10.3.
There is aqua-tk for OS X (which I used to make a screenshot). It
binds, as the name tells, directly to aqua, no X11 needed. So it
should be quite compliant (for example the menu bars automatically
are put in the top of the screen bar and not into the windows). It
uses native Aqua widgets, so I expect not too many problems with
the Aqua guidelines. (If you as the application author obey them :)
But I am not very experienced with Aqua, so if you notice something
that can be fixed from the Ltk side, please tell me.
Espen Vestre
> What about FreeBSD?
There's no specific FreeBSD version, but I've had positive reports
that the delivered applications I make for LW for linux work with
FreeBSD. They also work with OpenBSD.
--
(espen)
Camm Maguire
rp...@rpw3.org (Rob Warnock) writes:
> Paul F. Dietz <di...@dls.net> wrote:
> +---------------
> | Edward Jason Riedy wrote:
> | > I wasn't maligning the Altix line; they're damned impressive.
> | > I was noting that I'm not aware of an optimized, compiled Lisp
> | > for it. Am I wrong? I'd love to be wrong... (I don't have
> | > access to an Altix at the moment, but it's a target.)
> |
> | Is that one of the debian targets? gcl is compiled and runs on all
> | the debian target architectures.
> +---------------
>
> Hmmm... According to <URL:http://www.sgi.com/servers/altix/configs.html>
> they only support "SUSE LINUX Enterprise Server 8 + Service Pack 3", and
> that only with "SGI ProPack" added. Does GCL run on SuSE? In 64-bit mode?
>
From what I can see, these altix boxes are Itanium 2 Linux machines,
and gcl should compile out of the box on same, regardless of Linux
'flavor'. On Debian ia64, gcl provides the base for maxima, acl2, and
axiom, each of which pass all their own internal regression tests.
There are users of very large gcl images on multiprocessor Itanium
machines with a lot of memory in which more than 1 billion cons
elements can be allocated -- these users are primarily interested in
acl2 there. GCL also compiles out of the box on SuSE amd64 with the
same results as above, should you be wondering about SuSE.
You should note that gcl uses dlopen to load compiled object modules
on ia64, and hence should use a slightly different method for building
and saving images on top of it than is customary on i386. I can
provide details on request. amd64 on the other hand is just like x86
in this regard. Both are fully 64bit lisp systems.
> Also, does GCL support multiprocessors?
>
GCL is not currently multi-threaded, but does have a very good C
interface, making MPI implementations and the like quite easy. Check
out pargcl, an interface for using external MPI libraries to
communicate between gcl images across cluster nodes via tcp, or
between processors on the same machine using shared memory. There is
also a pvm interface at ftp://ftp.ma.utexas.edu/gcl. Lastly, it is
quite simple to write your own function calling 'fork()' and
communicating to the child via pipes via the gcl utility (clines
"...."), which has the same effect between lisp and C as the inlined
assembler directive __asm__ does between C and asm.
Take care,
>
> -Rob
>
> -----
> Rob Warnock <rp...@rpw3.org>
> 627 26th Avenue <URL:http://rpw3.org/>
> San Mateo, CA 94403 (650)572-2607
>
--
Camm Maguire ca...@enhanced.com
==========================================================================
"The earth is but one country, and mankind its citizens." -- Baha'u'llah
David Magda
> There's no specific FreeBSD version, but I've had positive reports
> that the delivered applications I make for LW for linux work with
> FreeBSD. They also work with OpenBSD.
Both of these have a Linux ABI emulation layer:
http://www.freebsd.org/cgi/man.cgi?query=linux
http://www.openbsd.org/cgi-bin/man.cgi?query=compat_linux
--
David Magda <dmagda at ee.ryerson.ca>, http://www.magda.ca/
Because the innovator has for enemies all those who have done well under
the old conditions, and lukewarm defenders in those who may do well
under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI
Espen Vestre
> Espen Vestre <espen@*do-not-spam-me*.vestre.net> writes:
>
>> There's no specific FreeBSD version, but I've had positive reports
>> that the delivered applications I make for LW for linux work with
>> FreeBSD. They also work with OpenBSD.
>
> Both of these have a Linux ABI emulation layer:
Yes, I forgot to mention that you need to have that installed.
--
(espen)
Luke Gorrie
> Luke Gorrie wrote:
>> Apologies if this has been hashed out before, but what are the
>> features that make the commercial Lisps so much better?
>
> In one word: complete integration.
>
> Editor, listener, debugger, inspector, various browsers all offered by
> one process which sees it all. So I can click on a stack frame in
> which function X gets invoked and hit CONTROL-ALT-. (jump to
> definition in source) and be investigating the source in about two
> seconds. And this pattern works no matter where I am and no matter
> which other navigational tool I want to kick off.
And this differs from SLIME how?
> One measure of the importance to this is that someday I will finally
> squeeze off an RFE to Franz: "Please, no full stops after identifiers
> in error messages." The problem is AllegroCL error output appearing in
> the listener, say:
>
> Division by zero in: KT-COMPUTE-AVERAGE.
>
> Normally I could click on the /output text/ and hit ctrl-alt-., but
> that little full stop at the end of the error message, while a great
> comfort i am sure to the former grammar teachers of Franz developers,
> prevents that.
Truly you are blessed to be able to nag somebody else to fix this for
you for money. We lost souls in the free software world would have to
resort to something drastic, like writing five lines of code to teach
the listener to treat . as punctuation instead of a symbol
constituent:
(add-hook 'slime-repl-mode-hook 'my-change-listener-symbol-syntax)
(defun my-change-listener-symbol-syntax ()
"Modify the syntax table to choice for the listener."
(set-syntax-table (copy-syntax-table))
(modify-syntax-entry ?\. "."))
Poor helpless us.
> Another thing is something like apropos being an interactive dialog in
> which one can reshape the query and see diffferent result sets as I
> type, not just the apropos output scrolling by in a listener.
It updates the result set after each keystroke? That sounds nice, we
might have to do that. Meanwhile `C-c C-a foo RET' is okay for
bringing up a nicely formatted and hyperlinked apropos listing in its
own scrollable/dismissable buffer for me.
> Well, OK: also support. When CMUCL or SBCL stagger to their feets on
> win32 will they come with installers? How about an automatic
> "update-cmucl" function which reaches out to common-lisp.net to
> download patches and apply them automatically?
In SLIME we have this command, it's called "cvs update". You can use
tags to choose whether you get the current sources or the latest
stable version. We handle any recompilation needed automatically. If
you put it in a shell script called 'update-slime' it would be
indistinguishable from what you describe as far as I can see.
Kenny, I'm all in favour of over-the-top enthusiastic outbursts, but
for heaven's sake do your homework before piling too much shit on your
fellow hackers. If you dislike the free environments because of
personal tastes then that's all you need to say to make your point.
Cheers,
Luke
Luke Gorrie
> I'm all in favour of over-the-top enthusiastic outbursts, but for
> heaven's sake do your homework before piling too much shit on your
> fellow hackers.
I should take my own advice. Apologies for raising the temperature in
here.
Happy midsummer,
Luke
Raymond Wiker
Midsummer here in Norway could do with a bit of
temperature-raising, so I thank you for your efforts :-)
Kenny Tilton
Luke Gorrie wrote:
> Kenny Tilton <kti...@nyc.rr.com> writes:
>
>
>>Luke Gorrie wrote:
>>
>>>Apologies if this has been hashed out before, but what are the
>>>features that make the commercial Lisps so much better?
>>
>>In one word: complete integration.
>>
>>Editor, listener, debugger, inspector, various browsers all offered by
>>one process which sees it all. So I can click on a stack frame in
>>which function X gets invoked and hit CONTROL-ALT-. (jump to
>>definition in source) and be investigating the source in about two
>>seconds. And this pattern works no matter where I am and no matter
>>which other navigational tool I want to kick off.
>
>
> And this differs from SLIME how?
I have heard great things about SLIME, and if/when I make my homecoming
to Mac OS X I will give it a try on SBCL.
>
>
>>One measure of the importance to this is that someday I will finally
>>squeeze off an RFE to Franz: "Please, no full stops after identifiers
>>in error messages." The problem is AllegroCL error output appearing in
>>the listener, say:
>>
>> Division by zero in: KT-COMPUTE-AVERAGE.
>>
>>Normally I could click on the /output text/ and hit ctrl-alt-., but
>>that little full stop at the end of the error message, while a great
>>comfort i am sure to the former grammar teachers of Franz developers,
>>prevents that.
>
>
> Truly you are blessed to be able to nag somebody else to fix this for
> you for money. We lost souls in the free software world would have to
> resort to something drastic, like writing five lines of code to teach
> the listener to treat . as punctuation instead of a symbol
> constituent:
>
> (add-hook 'slime-repl-mode-hook 'my-change-listener-symbol-syntax)
>
> (defun my-change-listener-symbol-syntax ()
> "Modify the syntax table to choice for the listener."
> (set-syntax-table (copy-syntax-table))
> (modify-syntax-entry ?\. "."))
>
> Poor helpless us.
Touche! But you digress. We are discussing the I in IDE, not the agreed
advantages of open source.
>
>
>>Another thing is something like apropos being an interactive dialog in
>>which one can reshape the query and see diffferent result sets as I
>>type, not just the apropos output scrolling by in a listener.
>
>
> It updates the result set after each keystroke? That sounds nice, we
> might have to do that. Meanwhile `C-c C-a foo RET' is okay for
> bringing up a nicely formatted and hyperlinked apropos listing in its
> own scrollable/dismissable buffer for me.
No, sorry, I misspoke. The dialog waits for you to hit enter. But then
as I play with filters the result set gets updated automatically.
In case any of these features interest you: one filter is "exported
symbols only". A radio group offers: all, functions, variables, or
classes only. a check box selects "all packages", and when off a pop up
of packages becomes enabled to specify which. The listing shows symbol,
package, and then flags for whether the symbol is exported, is bound to
a function, a setf, a variable, or a class (one col each).
Maybe I should just recreate it using Celtic and then you can see what I
am talking about. :)
>
>>Well, OK: also support. When CMUCL or SBCL stagger to their feets on
>>win32 will they come with installers? How about an automatic
>>"update-cmucl" function which reaches out to common-lisp.net to
>>download patches and apply them automatically?
>
>
> In SLIME we have this command, it's called "cvs update".
You *n*x people really have no idea how much you have learned about how
many tools. The Cells II release is a zip on ftp simply because i could
not get WinCVS to do anything. I figured it out once, but now I cannot
remember what I did. This is a very bad sign for the tool.
What I want is to click on "Install SBCL+Emavs+SLIME" on some web site
and click ok and next and agreed a dozen times and be presented with,
well, a text editor? Will it have a Slime menu, at least, so I can see
what is available in the way of tools?
Remember, I am an application programmer. I can have Lispworks up in
(well, I do not know, how automated is an LW install on Linux? If it
sucks, that is just another knock on free software.)
You can use
> tags to choose whether you get the current sources or the latest
> stable version. We handle any recompilation needed automatically. If
> you put it in a shell script called 'update-slime' it would be
> indistinguishable from what you describe as far as I can see.
>
> Kenny, I'm all in favour of over-the-top enthusiastic outbursts, but
> for heaven's sake do your homework before piling too much shit on your
> fellow hackers. If you dislike the free environments because of
> personal tastes then that's all you need to say to make your point.
No, I would not have said that if I had not heard from a respected free
Lisp user and developer that integration in free Lisps is not what it is
in environments where the editor and compiler and everything are all
part of the same application.
But that was a year ago and I gather Slime has made great strides since
then, which nicely brings us back on topic: If Slime really is so great,
what happened to the newbie (forgotten) to whom I responded with my
flame on free Lisps? The fellow seems to have done his best to make them
work for him.
kt
Matthew Danish
> You *n*x people really have no idea how much you have learned about how
> many tools. The Cells II release is a zip on ftp simply because i could
> not get WinCVS to do anything. I figured it out once, but now I cannot
> remember what I did. This is a very bad sign for the tool.
I can never get WinCVS to do anything either, and I'm fairly experienced
with CVS. I advise not using it, and just typing "cvs update" from the
command line in the source directory; it's very simple. WinCVS seems to
be an excellent example of GUI-gone-wrong.
> What I want is to click on "Install SBCL+Emavs+SLIME" on some web site
> and click ok and next and agreed a dozen times and be presented with,
> well, a text editor? Will it have a Slime menu, at least, so I can see
> what is available in the way of tools?
Does this count: http://www.common-lisp.net/project/lispbox ?
(Sorry for the rather sparse look, I need to restore files from back-up)
> Remember, I am an application programmer. I can have Lispworks up in
> (well, I do not know, how automated is an LW install on Linux? If it
> sucks, that is just another knock on free software.)
It is very simple. Just a matter of unpacking an archive.
William Bland
> But that was a year ago and I gather Slime has made great strides since
> then, which nicely brings us back on topic: If Slime really is so great,
> what happened to the newbie (forgotten) to whom I responded with my
> flame on free Lisps? The fellow seems to have done his best to make them
> work for him.
I think I've said this before on this newsgroup, but I'll say it again:
SLIME is *the* reason that I stopped being scared of Common Lisp and made
the leap from Scheme. It's good. Try it.
Cheers,
Bill.
--
Dr. William Bland.
It would not be too unfair to any language to refer to Java as a
stripped down Lisp or Smalltalk with a C syntax. (Ken Anderson).
Paolo Amoroso
> You *n*x people really have no idea how much you have learned about
> how many tools. The Cells II release is a zip on ftp simply because i
> could not get WinCVS to do anything. I figured it out once, but now I
No big deal, just use Emacs' VC (Version Control) mode :)
Björn Lindberg
> But that was a year ago and I gather Slime has made great strides
> since then, which nicely brings us back on topic: If Slime really is
> so great, what happened to the newbie (forgotten) to whom I responded
> with my flame on free Lisps? The fellow seems to have done his best to
> make them work for him.
If the Allegro IDE is so great, how come several people in this thread
said they can't stand it? Perhaps different tools suits different
people.
Björn
Thomas F. Burdick
> Kenny Tilton <kti...@nyc.rr.com> writes:
>
> > You *n*x people really have no idea how much you have learned about
> > how many tools. The Cells II release is a zip on ftp simply because i
> > could not get WinCVS to do anything. I figured it out once, but now I
>
> No big deal, just use Emacs' VC (Version Control) mode :)
It only helps some, pcl-cvs makes up most of the difference, but you
still need to consult wizzards for things like imports.
Kenny Tilton
Björn Lindberg wrote:
> Kenny Tilton <kti...@nyc.rr.com> writes:
>
>
>>But that was a year ago and I gather Slime has made great strides
>>since then, which nicely brings us back on topic: If Slime really is
>>so great, what happened to the newbie (forgotten) to whom I responded
>>with my flame on free Lisps? The fellow seems to have done his best to
>>make them work for him.
>
>
> If the Allegro IDE is so great, how come several people in this thread
> said they can't stand it?
I heard them saying they did not like the /editor/. Which only makes
sense, since they were hardcore emacs users and we are all programmed by
our editors to understand nothing else. But of course a little
familiarity and they would be as productive as ever.
Perhaps different tools suits different
> people.
Why you ol' relativist, you! No, the reason I spout the crap I do here
on c.l.l. is that I always ask /fans/ of XXX before trashing it. ie,
some tools just plain break our balls. My sample is too small, but then
this is Usenet, not a frickin refereed journal.
Hey, how about celtk? Then no onewill know how to pronounce it? Almost
looks like Celtic if you cross your eyes, and then we have a most
excellent compression of Cells-LTk-Tk. celltk? some camelcase: ceLTk?
nah. i'll sleep on celtk. cvs release coming soon, thx to thems that
know better.
Brian Downing
Kenny Tilton <kti...@nyc.rr.com> wrote:
> I heard them saying they did not like the /editor/. Which only makes
> sense, since they were hardcore emacs users and we are all programmed by
> our editors to understand nothing else. But of course a little
> familiarity and they would be as productive as ever.
"I heard them saying they did not like the /language/. Which only makes
sense, since they were hardcore CL users and we are all programmed by
our languages to understand nothing else. But of course a little
familiarity in Java and they would be as productive as ever."
:-), mostly. It is possible, however, that GNU Emacs is simply a more
powerful/expressive editor than the built-in Allegro CL one.
-bcd
--
*** Brian Downing <bdowning at lavos dot net>
sir...@fisica.unige.it
Hi Kenny,
>>>>> "Kenny" == Kenny Tilton <kti...@nyc.rr.com> writes:
Kenny> You *n*x people really have no idea how much you have
Kenny> learned about how many tools. The Cells II release is a zip
Kenny> on ftp simply because i could not get WinCVS to do
Kenny> anything. I figured it out once, but now I cannot remember
Kenny> what I did. This is a very bad sign for the tool.
If you don't like WinCVS, you should probably try TortoiseCVS. It's
completely integrated into the windows explorer. Very nice. Anyway, i
don't think CVS alone would be a good installing tool for win32. It
isn't very good neither for unixes: as far as I know, none of the unix
flavours use cvs as an installing tool. With some BSD you can use it
for upgrading the ports tree (i.e. the tree of makefiles used to install
SW), but further tools are used in order to install the software (the
most basic being make).
Bye,
e.
--
Enrico Sirola <sir...@fisica.unige.it>
Espen Vestre
> You *n*x people really have no idea how much you have learned about
> how many tools.
This is a point frequently forgotten by unix geeks. You have to
eat, breathe and drink unix to really appreciate it, if you use
it too infrequently, it will always seem almost inprenetable
(however, nowadays you can profit from unix without really under-
standing it, with os x or linux gui environment).
> (well, I do not know, how automated is an LW install on Linux? If it
> sucks, that is just another knock on free software.)
The install itself is very simple, even if you use the tar'ed version
and not the redhat-rpm. The only potential problem is getting the
right motif version (I wish Xanalys could move to a more modern gui
toolkit!)
> No, I would not have said that if I had not heard from a respected
> free Lisp user and developer that integration in free Lisps is not
> what it is in environments where the editor and compiler and
> everything are all part of the same application.
I agree that the Emacs-and-lisp-as-separate-worlds solution isn't
ideal, I love the way FRED is integrated into MCL, and I prefer
to use the integrated editor when using LispWorks (although it's
a bit more opaque than FRED, not quites as easy to program).
However, my colleagues both where I work now and at my previous
company are divided in how they prefer to use LispWorks: Some think
that integration with The Real Thing (Emacs) is more important than
closeness to the lisp enviroment, others (such as I) think the
opposite. All in all, YMMV, and there are good reasons for choosing
either solution.
--
(espen)
Jock Cooper
If someone can show me how to do an electric-buffer-menu,
split/unsplit windows (horizontally and vertically), visit/save/close
multiple files, isearch, named kill buffers, simple macros, operate on
rectangles, seemingly unlimited undos, checkout/in, and so on, all
without my hands leaving the keys, I'll be happy to give ACL's editor
a chance. In fact I don't want my hands leaving the home row--so no
arrow keys, or pageup/dn etc.
When I get into that editor I feel like I'm in notepad.
David Steuber
> But that was a year ago and I gather Slime has made great strides since
> then, which nicely brings us back on topic: If Slime really is so great,
> what happened to the newbie (forgotten) to whom I responded with my
> flame on free Lisps? The fellow seems to have done his best to make them
> work for him.
I've got Lispworks Personal edition sitting on my dock and have yet to
fire it up.
I find it interesting that there are people using non-free (as in not
open source) Lisps with SLIME. LW and ACL being notable in that
respect. Considering that IDEs are available for at least those two
Lisps, one has to wonder why people would use SLIME in preference to
the bundled IDE. The only reason that I can think of is that perhaps
they prefere the one over the other even though SLIME has to use IPC
and some very clever Lisp hackery to provide a common look and feel
(as far as is possible) among different Lisps on different platforms.
I'm very glad for SLIME. Without it I probably would not have gotten
as far as I have.
The commercial Lisps may be technically better (or not). I still
think that it is important to have a few good free Lisp
implementations to choose from and experiment with. Not everyone is
willing to pay for a Lisp just to learn it (although LW and ACL do
have trial versions which I think is a very good thing). Languages
like Python and Java give people certain expectations of features in a
language (reasonable or not). I believe that Lisp (meaning Common
Lisp) is the superior language. I think that "missing" features
reduce the chances of Lisp capturing back mind share. This has the
further result of reducing the developer base. This also impacts
commercial Lisps.
If I never met Lisp, I would not realize that languages such as Java
are a technological culdesac. I would probably be a happier person
for it.
Lisp has also challenged my conceptions of what programming is all
about.
Christopher C. Stacy
David> I find it interesting that there are people using non-free (as in not
David> open source) Lisps with SLIME. LW and ACL being notable in that respect.
David> [...] The only reason that I can think of is that perhaps [...]
David> to provide a common look and feel (as far as is possible)
David> among different Lisps on different platforms.
The main reason I can think of is that the editors in the commercial
IDEs are slightly different (in some non-customizable way) from GNU Emacs,
or are missing some features that they like and that it's too much
trouble to add and maintain those features.
That's what Jock Cooper's message seems to be, too.
Peter Herth
> I find it interesting that there are people using non-free (as in not
> open source) Lisps with SLIME. LW and ACL being notable in that
> respect. Considering that IDEs are available for at least those two
> Lisps, one has to wonder why people would use SLIME in preference to
> the bundled IDE. The only reason that I can think of is that perhaps
> they prefere the one over the other even though SLIME has to use IPC
> and some very clever Lisp hackery to provide a common look and feel
> (as far as is possible) among different Lisps on different platforms.
There are several reasons to prefer emacs over an IDE of whatever
making and especially using SLIME. These are not religious :) but
at least for me very practical ones. For one, SLIME has some nice
features, they lack (or perhaps I just have not found in my trials
with the LW personal edition). But there is one fundamental
reason why emacs is "better" than any IDE, and that is support for
more than one programming language. If you do only Lisp or only
Java, any of the IDEs may be very good, though the emacs editing
modes are quite sophisticated too. But during a normal day I edit
code in Lisp, Java, Python, Scheme and LaTeX. So even if some
flashy features are missing in the correspondant emacs mode, I
do not only have not to open a custom editing application for
each file, also it is a huge advantage to have a very consistant
editing interface across several languages. So in my eyes, for
all those who work with several programming languages, emacs
is more practical than custom IDEs.
Hannah Schroeter
Peter Herth <he...@netcologne.de> wrote:
>> Actually, a pet peeve I have about Java is the fact that you call new
>> to create an object but there is no delete to balance it like with
>> C++. If you have automatic memory management, why do you need new?
>> It should be implicit. I suppose one might have the same complaint
>> about make-array or make-hash-table, but somehow I don't see anything
>> wrong with those functions.
>Well, with an implicit new, any List structure would be in infinite
>size if you have:
>class Cons
>{
> Object car;
> Cons cdr; // <= if this allocated a Cons object by default, this would mean
>trouble :)
>}
Yep.
But why do you have to write Cons foo = new Cons instead of
just Cons foo = Cons() or so? Or with type inference (or dynamic
typing): foo = Cons(), or even foo = Cons(a, b), then it suddenly
begins to look like that beautiful
(let ((foo (cons a b)))
...)
thing.
You see, no "new" in there. No (new 'cons a b).
But then, you could of course, in Java, put a class just above
Object, having a
static Cons cons(Object a, Cons b)
{
Cons c = new Cons();
c.car = a;
c.cdr = b;
return (c);
}
in there. Inherit that class either directly or indirectly for
every other class you define.
>[...]
Kind regards,
Hannah.
Peter Herth
> But why do you have to write Cons foo = new Cons instead of
> just Cons foo = Cons() or so? Or with type inference (or dynamic
> typing): foo = Cons(), or even foo = Cons(a, b), then it suddenly
> begins to look like that beautiful
> (let ((foo (cons a b)))
> ...)
> thing.
>
> You see, no "new" in there. No (new 'cons a b).
Yes of course you can make it more beautiful than Java chose to
do but Cons() is still an explicit call and I understood the
original poster in the way that no explicit call ist needed and
that creates some problems with list-like structures.
Rahul Jain
> You *n*x people really have no idea how much you have learned about how
> many tools. The Cells II release is a zip on ftp simply because i could
> not get WinCVS to do anything. I figured it out once, but now I cannot
> remember what I did. This is a very bad sign for the tool.
So use a client app that works. Like the normal CVS tool.
> What I want is to click on "Install SBCL+Emavs+SLIME" on some web site
> and click ok and next and agreed a dozen times and be presented with,
> well, a text editor? Will it have a Slime menu, at least, so I can see
> what is available in the way of tools?
Too bad everyone who uses windows (or anything but debian for that
matter) doesn't want to create a tool that lets you do this. Why do you
people keep bringing up this topic when you're violently against doing
anything about it?
> Remember, I am an application programmer. I can have Lispworks up in
> (well, I do not know, how automated is an LW install on Linux? If it
> sucks, that is just another knock on free software.)
Why would the fact that installing LW, ACL, etc is more difficult than
installing cmucl, sbcl, etc be a knock on free software?
> But that was a year ago and I gather Slime has made great strides since
> then, which nicely brings us back on topic: If Slime really is so great,
> what happened to the newbie (forgotten) to whom I responded with my
> flame on free Lisps? The fellow seems to have done his best to make them
> work for him.
Or maybe he's against using the only system where making these things
work is easy. It irritates me to have to reiterate this, but sadly, no
one is FOR making these things work easily on other systems. Oh, and
don't pretend that this is a free lisp vs. commercial lisp thing.
Installing LW on windows isn't as easy as installing cmucl on debian.
In any case, I _do_ agree (it's the end of the world! ;) with your
sentiment about IDEs being integrated into the app. Luke and friends
have done quite a bit of work in trying to get the emacs<->CL
communication to work well, and it seems to be rather different for
different CLs.
--
Rahul Jain
rj...@nyct.net
Professional Software Developer, Amateur Quantum Mechanicist
Kenny Tilton
Rahul Jain wrote:
> Kenny Tilton <kti...@nyc.rr.com> writes:
>
>
>>You *n*x people really have no idea how much you have learned about how
>>many tools. The Cells II release is a zip on ftp simply because i could
>>not get WinCVS to do anything. I figured it out once, but now I cannot
>>remember what I did. This is a very bad sign for the tool.
>
>
> So use a client app that works. Like the normal CVS tool.
Raw CVS? That's a good one.
Fortunately a better-informed respondent (thanks) recommended
TortoiseCVS and I have now moved Cells II, Cello, Celtic, Cloucell, and
a nascent Visual Apropos into cvs, along with cl-opengl, cl-openal,
cl-magick, and cl-ftgl. Formal announcement after Mr. Burdick patches
Cells II for cmucl compatibility.
Alain Picard
> I find it interesting that there are people using non-free (as in not
> open source) Lisps with SLIME. LW and ACL being notable in that
> respect. Considering that IDEs are available for at least those two
> Lisps, one has to wonder why people would use SLIME in preference to
> the bundled IDE.
It's really very simple. There exists a category of users (I'm one
of them) for which getting out of emacs is almost physically painful.
It's hard to explain, but, in emacs, I just *think* what I want to
happen, and it happens. In every other software I've ever used,
it's click here, point there, read this manual, blah blah blah, and
by the time one figures it out, you've forgotten what you were actually
trying to do. So, Mail, News, CVS, dired, shell, (and yes, lisp) all
work better within emacs: to-ge-ther.
For those who think Notepad is a good enough editor, this is a difficult
concept to understand.
Björn Lindberg
> Björn Lindberg wrote:
> > Kenny Tilton <kti...@nyc.rr.com> writes:
> >
> >>But that was a year ago and I gather Slime has made great strides
> >>since then, which nicely brings us back on topic: If Slime really is
> >>so great, what happened to the newbie (forgotten) to whom I responded
> >>with my flame on free Lisps? The fellow seems to have done his best to
> >>make them work for him.
> > If the Allegro IDE is so great, how come several people in this
> > thread
> > said they can't stand it?
>
> I heard them saying they did not like the /editor/. Which only makes
> sense, since they were hardcore emacs users and we are all programmed
> by our editors to understand nothing else. But of course a little
> familiarity and they would be as productive as ever.
The biggest difference I saw in how they describe their working style
as compared to yours, is that they want to be able to exclusively use
the keyboard, whereas you seem content with using the mouse a lot as
well as the keyboard. Is the Allegro IDE as convenient to use with
only the keyboard as Emacs is? (I have not used the Allegro IDE
myself.)
Björn
André Thieme
Btw, switching between mouse and keyboard can also be very easy:
http://www.fingerworks.com/
No time loss with the TouchStream LP.
André
--
Kenny Tilton
Yes, it has key chords defined for everything. It also has a feature I
prefer to emacs buffers: a tab control which shows one source at a time,
but with a row of tabs (buttons) across the top showing what else I have
open. I am so comfortable with the mouse that I just discovered (thx to
this digression (no one said emacs was not a great editor, nor that
programmers fall twitching to the ground if exposed for more than three
minutes to a different editor)) it has key chords for moving left/right
thru those tabs. It also has chords for returning to prior tab or
secondmost prior (penultimate?) tab. I listed in an earlier message all
the chords with their functions, no one attacked that. We did hear about
split windows -- I can open a second editor-cum-tab-control to see
source side by side, and I have done that twice in five years.
But again, I am sure I would love editing with emacs, just not enough to
give up the integration with all the tools, because ACL really does have
a fine built-in editor. Comparisons to Notepad are, er, um, misleading.
It just says "this is how I feel using this unfamiliar editor". No
doubt. I feel paralyzed when using emacs or LW's Hemlock workalike
(except that as I have said, I use ACL in "emacs mode" so I do OK if I
stick to the few things I know are emacs-like (from my Fred days).
Maybe Slime changes all this -- maybe at the next lisp-nyc meeting rahul
or some other slimeball can give a demo. of course by then he will be
able to run the Celtic sample code for visual-apropos (except he will
have to rewrite it without Cells or be kicked permanently from Lisp IRC.)
:)
kt
ps. I am rooting for Slime. If Slime succeeds, I can move back to Mac OS
X and use emacs+slime+sbcl (or openmcl? other?). I would /love/ to leave
this XP garbage behind.
Rahul Jain
> It also has a feature I prefer to emacs buffers: a tab control which
> shows one source at a time, but with a row of tabs (buttons) across
> the top showing what else I have open.
Shocking. Emacs NEVER had a feature like that. I'm ALWAYS missing those
tabs at the top of my screen (ok, they're on the right side at work)
when I'm using Emacs. Especially tabs that change to show only the
buffers in the same mode as the one you're looking at now, but with the
option to filter them or not in any way you choose. Yup, the Allegro IDE
really beats Emacs on that one.
> Maybe Slime changes all this -- maybe at the next lisp-nyc meeting rahul
> or some other slimeball can give a demo. of course by then he will be
> able to run the Celtic sample code for visual-apropos (except he will
> have to rewrite it without Cells or be kicked permanently from Lisp IRC.)
Oh, and where did you get the idea that I use SLIME? (It does have a
rather fancy apropos feature, since you mention that... or was that the
McCLIM Listener...)
:)
But seriously, your idiotic flamebaiting of those who don't care for
your reinvention of the 15-sided, concave wheel is starting to lose its
entertainment value.
Raistlin Magere
"Alain Picard" <Alain....@memetrics.com> wrote in message
news:87pt7lu...@memetrics.com...
> da...@david-steuber.com (David Steuber) writes:
>
> It's hard to explain, but, in emacs, I just *think* what I want to
> happen, and it happens. In every other software I've ever used,
> it's click here, point there, read this manual, blah blah blah, and
> by the time one figures it out, you've forgotten what you were actually
> trying to do. So, Mail, News, CVS, dired, shell, (and yes, lisp) all
> work better within emacs: to-ge-ther.
>
> For those who think Notepad is a good enough editor, this is a difficult
> concept to understand.
Ok, what is the best way to get started to emacs (or should I try xemacs?).
I am aware that emacs comes with a manual but at this moment of time I don't
quite have the time (sorry for the word repetion) to read a manual in detail
so I would like to know if there's a quick and dirty "emacs for dummies"
guide somewhere.
My needs are simple as I only program in Lisp (Matlab now and again) and
write in LaTeX, so far I have been using the ACL5.01 editor and WinEdt for
my needs and I must say that I have found them v.good (I also discovered acl
keychords only recently - i.e. in the past 5,6 months and they have greatly
improved my life) furthermore as I have never really figured out how to use
the gui creator of acl I will not really miss that part of integration.
I am actually quite happy with the tools I am using at the moment but I am
happy to learn to use a more powerful tool.
Thanks for your help,
raistlin