Avast mateys!
I have come to the shocking realisation that we actually have USERS!
Good heavens.. we haven't really prepared for this. I feel like the proverbial guy who forgot he was hosting a party and only had light beer when the guests arrived.
What we need is a way for new users to find out how to use SLIME -- what some would call a manual. We need the information to be easy to update and get feedback on and for changes to be easy to follow. Sounds like a job for Cliki! So I've started a couple of new pages there:
http://www.cliki.net/SLIME-HOWTO The mundane stuff - how to download and install.
http://www.cliki.net/SLIME%20Features What features we have and how to use them.
I have only made a start on the Features page. It would be great if newcomers to SLIME post little notes pointing out important information that's missing, and if seasoned users and hackers help them out and write a few words about their favourite features.
What else can we do to help people use SLIME, beyond hacking?
Cheers, Luke
"Luke" == Luke Gorrie luke@bluetail.com:
Luke> Avast mateys! I have come to the shocking realisation that we Luke> actually have USERS! [...]
"Build it, and they will come..."
Seriously, SLIME has already gotten to the point where it's more useful than ILisp for me, so I've just switched over. Yes, it has its problems and is a little too explosive at times, but it does some things very nicely and actually has a future (which matters), being actively developed.
Luke> What else can we do to help people use SLIME, beyond hacking?
Listen to the people :-) -- I'm trying to suggest features or functionality from time to time, so do others. Polish the details and SLIME will achieve greatness and fame!
A good manual (like the Gnus one) would be a good addition, too, but perhaps it's too early for that, as the software still changes often.
Thanks for your efforts!
--J.
Jan Rychter jan@rychter.com writes:
Seriously, SLIME has already gotten to the point where it's more useful than ILisp for me,
I switched to Slime for awhile, but switched back to Ilisp for the moment for two reasons:
- stability. This is probably a red herring, since you can evidently just leave a Lisp up and running and reconnect after a crash; it's also transient, since Slime is improving by leaps and bounds (infinitely faster than Ilisp, one might say;).
- Usability details. The big one here is completion -- Ilisp has a couple of completion features that Slime lacks: filename completion in strings, and multiword expansion (e.g. "m-w-e<tab>" -> "multi-word-expansion"). Is there any reason not to pillage Ilisp for these nice touches?
If I get some free time over the holidays, I'll try to address these two issues with some patches, and get back to happily using Slime. It's got a lot of potential, but is still to young for someone who just needs to get his Lisp work done with as few hassles as possible.
/s
"Sean" == Sean O'Rourke sorourke@cs.ucsd.edu:
Sean> Jan Rychter jan@rychter.com writes:
Seriously, SLIME has already gotten to the point where it's more useful than ILisp for me,
Sean> I switched to Slime for awhile, but switched back to Ilisp for Sean> the moment for two reasons:
Sean> - stability. [...]
Yes, that's what I meant by "explosive". In particular, I've been bitten by SLIME's unwillingness to act gracefully when lisp decides to produce tons of output. I found most of the time I have no way to interrupt the process if it is producing output fast, which means I eventually have to kill XEmacs. Neither C-c C-c nor C-g in the REPL do anything.
Sean> - Usability details. The big one here is completion -- Ilisp has Sean> a couple of completion features that Slime lacks: filename Sean> completion in strings, and multiword expansion (e.g. "m-w-e<tab>" Sean> -> "multi-word-expansion"). Is there any reason not to pillage Sean> Ilisp for these nice touches?
Oh, yes. I miss the multiword expansion, too. I have to remind myself not to type m-v-b<TAB> every once in a while.
--J.
"Sean O'Rourke" sorourke@cs.ucsd.edu writes:
- stability. This is probably a red herring, since you can evidently just leave a Lisp up and running and reconnect after a crash; it's also transient, since Slime is improving by leaps and bounds (infinitely faster than Ilisp, one might say;).
Could you explain a bit what stability problems you see? And what combination of Emacs/Lisp are you using?
"Sean O'Rourke" sorourke@cs.ucsd.edu writes:
I switched to Slime for awhile, but switched back to Ilisp for the moment for two reasons:
- stability. This is probably a red herring, since you can evidently just leave a Lisp up and running and reconnect after a crash; it's also transient, since Slime is improving by leaps and bounds (infinitely faster than Ilisp, one might say;).
I hadn't realised we had stability-as-in-crash problems. We'll fix them.
- Usability details. The big one here is completion -- Ilisp has a couple of completion features that Slime lacks: filename completion in strings, and multiword expansion (e.g. "m-w-e<tab>" -> "multi-word-expansion"). Is there any reason not to pillage Ilisp for these nice touches?
Pillaging ideas, definitely. Pillaging code hasn't been a big success so far. In the cases I've tried, the ILISP code was really showing its age (way) too much.
Today might be an exception. I spent quite some time implementing a "m-w-e<tab>" feature, but the algorithm I cooked up turned out to completely suck, and it isn't useful. Bugger. :-)
Cheers, Luke
Hi,
I am not an expert of Ilisp nor Slime, but isn't it possible to make a completion command like "m<tab>" which works like the completion in the mini-buffer of Emacs.
It seems a very good idea to complete filenames, though difficult to port.
Isn't it possible to complete symbols against the result of an apropos funcall ? It would make the completion feature very useful and could be enhanced to automatically insert not unused packages like this:
For example, if I want to write start-server which is in package ASERVE-EXAMPLE (which must be loaded) but my file doesn't use package ASERVE-EXAMPLE for some reasons. If I would write :
- "s<tab>" then a specific Emacs buffer would appear with the list of all symbols beginning by "s" in any packages. For instance we should have something like :
SYMBOL-VALUE ASERVE-EXAMPLE:START-SERVER
- then if I would type "y<tab" the symbol would complete to "SYMBOL-VALUE". - then if I would type "t<tab" the symbol would complete to "ASERVE-EXAMPLE:START-SERVER".
If find very annoying a Lisp environnement than can't complete symbol that the lisp knows. As I am not very good at lisp (still beginner !) I don't know if this is easy of difficult to write, but I suppose a quick-and-dirty hack would be easy by parsing the apropos funcall ...
Cheers, Camille
On Nov 26, 2003, at 5:52 PM, Luke Gorrie wrote:
"Sean O'Rourke" sorourke@cs.ucsd.edu writes:
I switched to Slime for awhile, but switched back to Ilisp for the moment for two reasons:
- stability. This is probably a red herring, since you can evidently just leave a Lisp up and running and reconnect after a crash; it's also transient, since Slime is improving by leaps and bounds (infinitely faster than Ilisp, one might say;).
I hadn't realised we had stability-as-in-crash problems. We'll fix them.
- Usability details. The big one here is completion -- Ilisp has a couple of completion features that Slime lacks: filename completion in strings, and multiword expansion (e.g. "m-w-e<tab>" -> "multi-word-expansion"). Is there any reason not to pillage Ilisp for these nice touches?
Pillaging ideas, definitely. Pillaging code hasn't been a big success so far. In the cases I've tried, the ILISP code was really showing its age (way) too much.
Today might be an exception. I spent quite some time implementing a "m-w-e<tab>" feature, but the algorithm I cooked up turned out to completely suck, and it isn't useful. Bugger. :-)
Cheers, Luke
slime-devel site list slime-devel@common-lisp.net http://common-lisp.net/mailman/listinfo/slime-devel
Camille Troillard tuscland@mac.com writes:
I am not an expert of Ilisp nor Slime, but isn't it possible to make a completion command like "m<tab>" which works like the completion in the mini-buffer of Emacs.
Are you familiar with the completion that we (and ILISP) have today? It is essentially like this. You can press M-TAB (or C-c TAB) to try it, or just TAB in the *slime-repl* buffer.
We don't treat packages the way you mention in your example though. We do have another form of integrated apropos search, bound to "C-c C-a".
Cheers, Luke
Jan Rychter jan@rychter.com writes:
Luke> What else can we do to help people use SLIME, beyond hacking?
Listen to the people :-) -- I'm trying to suggest features or functionality from time to time, so do others.
These are always appreciated. If we seem slow to take them up, it's because we're working on suggestions coming from other sources. The influx of ideas and bug reports from the list + cliki + IRC is enough to occupy a small nation. :-)
Cheers, Luke
-
Camille Troillard -
Helmut Eller -
Jan Rychter -
Luke Gorrie -
Sean O'Rourke