#lisp - Mon 19 Feb 2007 between 00:11 and 00:45

NY Lost Funds



sohailhey, does cl-launch force sbcl to not read .sbclrc?
beachgood morning
zeeeeein asdf, what's the difference between site and system?
(when i asdf-install:install, it asks where i want to install - each of system/user has site and system)
sohailzeeeee, one is for all users, the other is just for you I think
zeeeeesohail: er, isn't that the diff btwn system and user? each of those has a further option - site vs system
sohailoh!
brainly-greendoes there exist somewhere a "tiny" common lisp that consists of a tiny core with some macros and everything else defined in terms of the core?
but that is a compliant common lisp
Zhivagohmm, we need recalcitrant lisp
brainly-greenrecalcitrant?
sohailwe need a good lisp that runs under java
rr--tiny and common - isn't that impossible
brainly-greenno
beachbrainly-green: probably not, and the reason is because of performance. While you *could* define CL on top of a tiny core, that wouldn't necessarily be fast.
brainly-greenonly the core would be tiny, everything else would be defined in terms of the core but of course would be pretty big
I know it wouldn't be fast
Zhivagohmm, not sure why it would necessarily be slow, either.
beachbrainly-green: well, since most people want something fast, probably nobody has attempted something like that. Though perhaps Lisp500 (was that the name?) is something like that.
brainly-greenbut it would be nice then to extend the core with macros allowing you to smoothly inline C or machine code
so that the core could be tiny, but the language would be both powerful and fast
beachbrainly-green: what is it that you are trying to accomplish here?
sohaila small common lisp?
brainly-greenwell, I'm thinking that if 99% of a language is implemented as a library of macros on top of a core, then the language could evolve easily
because someone writing a compiler only needs to implement the core, and anyone who wants a new feature just writes an extension
which is then subjected to the public opinion
sohailbrainly-green, thats how common lisp works as I understand it
AshyIsMebrainly-green, you want #scheme
Zhivagowell, CL doesn't really work like that.
brainly-greenyes but with common lisp you can't vote OUT features that people don't find as useful
Zhivagobut I think that you can take a leaf from the book of C, here.
C considers all objects to be byte sequences.
brainly-greenbecause everything is implementation dependent
Zhivagoif you considered all objects to be word sequences in lisp, then it might be simpler.
pkhuongbrainly-green: there are only 25 special operators, and some of them are macro-expressible in terms of a smaller set.
Zhivagothen for speed, you could consider an object to be a word sequence with an optional byte sequence tail on it.
sohailZhivago, I know you're a genius and all but what does that have to do with it
pkhuongThat's the metric that matters when you want to write extensions.
Zhivagoat that point, it should become straightfoward to implement things like specialized arrays at a library level.
sohail: what does what have to do with what?

Page: 2 9 16 23 30 37 

IrcArchive

NY Lost Funds