Re:Learning Scheme in Emacs
by djcb on Sunday September 11, 2011 @06:32AM
(#2983580)
I think the first step is to leverage Guile as a 'better elisp interpreter'. A lot of progress has been made in that direction, lately. Now, if it will ever replace the existing elisp interpreter depends on a lot of things, technical and non-technical. Guile could (potentially) be significantly faster, and stuff like threads, FFI may be nice as well. Anyway, I wouldn't hold my breath :)
Re:Learning Scheme in Emacs
by alnayyir on Sunday September 11, 2011 @03:14AM
(#2983405)
As an avid elisper:
No. Just...no.
The lexical scope doesn't make sense for the kind of scripting done in Emacs and having a unified language (unlike vim with many people bolting their preferred language onto it) for the most part is a huge boon.
90% of the people I see rooting for this don't actually write anything for Emacs.
Re:Learning Scheme in Emacs
by muuh-gnu on Sunday September 11, 2011 @08:53AM
(#2983707)
As I understood it, the goal is not to simply allow Guile alongside Elisp, with the intent to confuse users, but to eventually replace Elisp completely. Guile devs are claiming that their Elisp implementation running on top of guile provides several substantial benefits over Emacs own C-implementation of elisp:
* higher speed
* native threads
* lexical scope
* ffi
* module system
* goops, guiles object system
Emacs is a key GNU project, but to the detriment of the overall project philosophy, it uses an extension language no other GNU project uses and never will use. Elisp may be working for emacs, but it is isolating it from all other apps, since nobody else uses Elisp. The long term goal of GNU is to "emacsify" the whole system, i.e. to enable emacs and other apps to talk to each other over guile es their universal scripting interface.
Re:Learning Scheme in Emacs
by muuh-gnu on Sunday September 11, 2011 @11:29AM
(#2983936)
> You people don't even
I'm not "you people", I'm not affiliated with Guile in any way. However, you are wrecking your brains over something without being fully informed. It is not and never has been intended to _rewrite_ anywthing written in Elisp. Thats why Elisp itself has been to a large part already rewritten in Guile, and it is planned to gut emacs and replace the current C-Elisp engine with the new Guile-Elisp engine. Whether this will or will not work out as intended is a completely different story. But the fact remains that Elisp, in the grand scheme of things, is an isolated island, and as such does not fit into the overall design goals of GNU stated over 20 yrs ago, so there are people working on this. Even Google sponsored two GSoC projects for two years to rewrite Elisp in Scheme.
Re:Learning Scheme in Emacs
by artlogic on Sunday September 11, 2011 @11:42AM
(#2983957)
Could you expand on "lexical scope is unnecessary and potentially even a bad idea"? I'm really curious given that fact that almost all the languages I use support lexical and not dynamic scope.
Re:Learning Scheme in Emacs
by abhiyerra on Sunday September 11, 2011 @04:23AM
(#2983470)
I agree with this sentiment. I like that Emacs has one language for its plugin and extension architecture as opposed to Vim's duck tape system of whatever language you want.
Elisp may not be the best at everything but as an editor scripting language it's quite great.
Re:Learning Scheme in Emacs
by djcb on Sunday September 11, 2011 @06:16AM
(#2983573)
Hmm, I usually use guile[1] + geiser[2] in emacs, which works quite well for me. There's even an info-version[3] of SICP, so all the fun can happen within emacs.
Re:Learning Scheme in Emacs
by gclaramunt on Sunday September 11, 2011 @01:00AM
(#2983259)
I was going through "The Little Schemer" and reached the point where pen & paper wasn't enough. Installed MIT Scheme but quickly got frustrated by Emacs. Tried DrScheme/DrRacket and it was awesome. Highly recommended
http://www.red-bean.com/guile/notes/emacs-lisp.html
No. Just...no.
The lexical scope doesn't make sense for the kind of scripting done in Emacs and having a unified language (unlike vim with many people bolting their preferred language onto it) for the most part is a huge boon.
90% of the people I see rooting for this don't actually write anything for Emacs.
* higher speed
* native threads
* lexical scope
* ffi
* module system
* goops, guiles object system
Emacs is a key GNU project, but to the detriment of the overall project philosophy, it uses an extension language no other GNU project uses and never will use. Elisp may be working for emacs, but it is isolating it from all other apps, since nobody else uses Elisp. The long term goal of GNU is to "emacsify" the whole system, i.e. to enable emacs and other apps to talk to each other over guile es their universal scripting interface.
It's. Not. Worth. It.
You people don't even listen when we try to explain that lexical scope is unnecessary and potentially even a bad idea.
Threads? Come on. :\
I'm not "you people", I'm not affiliated with Guile in any way. However, you are wrecking your brains over something without being fully informed. It is not and never has been intended to _rewrite_ anywthing written in Elisp. Thats why Elisp itself has been to a large part already rewritten in Guile, and it is planned to gut emacs and replace the current C-Elisp engine with the new Guile-Elisp engine. Whether this will or will not work out as intended is a completely different story. But the fact remains that Elisp, in the grand scheme of things, is an isolated island, and as such does not fit into the overall design goals of GNU stated over 20 yrs ago, so there are people working on this. Even Google sponsored two GSoC projects for two years to rewrite Elisp in Scheme.
See:
* http://wingolog.org/archives/2009/01/07/a-brief-history-of-g...
* http://wingolog.org/archives/2010/04/02/recent-developments-...
* http://wingolog.org/archives/2011/08/30/the-gnu-extension-la...
http://www.gnu.org/software/emacs/emacs-paper.html#SEC17
The docs and the man himself say it best.
Elisp may not be the best at everything but as an editor scripting language it's quite great.
[1] http://www.gnu.org/s/guile/ [2] https://savannah.nongnu.org/projects/geiser/ [3] http://www.neilvandyke.org/sicp-texi/