freenode/#lisp - IRC Chatlog
Search
12:45:06
ecraven
Xach: ah, there's only two parts so far? This seems like a book I would gladly pay for, but it hasn't been written yet :-/
13:09:27
Colleen
loke: drmeister said 8 hours, 52 minutes ago: The dev branch built - I pushed it to the master branch
13:11:17
drmeister
You may want to try it out - but it needs a special build of boehm and a little more setup than usual.
13:12:26
drmeister
Yes - we have not integrated the boehm build system into clasp - it's a bit of a PITA
13:13:19
drmeister
It's still "experimental" - but it's been working for weeks in several peoples hands.
13:14:19
drmeister
If you have more than ~8 cores the build becomes dominated by other linear processes.
13:16:10
_death
ecraven: I think a very strong point for lisp is that it's not just a language, but also a "living" system.. when developing you simultaneously evolve the lisp language and the lisp process
13:16:56
ecraven
_death: I've been working with SLIME a lot with different Schemes (implementing an r7rs swank), and I've come to found out that while I prefer the *language* Scheme, it just cannot compare in any way to the *system* Common Lisp
13:17:24
ecraven
I haven't found any Scheme that comes even remotely close to debugging, introspection and mutation of running systems in common lisp
13:17:50
ecraven
I am very much a fan of "world" systems, and it seems every Scheme implementation has mostly just given up on that :-/
13:18:38
_death
just today I put back a screencast I made a while ago.. it's a bit long, but may be entertaining to "advanced" lispers.. https://adeht.org/casts/new-project.html
13:18:53
ecraven
loke: years ago, I find it interesting, but not something I would want to actually use the way I use lisp now
13:19:01
loke
ecraven: Smalltalk is nothing like Lisp, but has the same system-based development model. If you try it, and you like the model there, ten that suggests that that is indeed the aspect you like.
13:19:20
ecraven
loke: oh, I like that aspect of smalltalk, I don't like all the visual things it necessitates
13:19:23
loke
ecraven: All of those languages are more or less system-based in the way you described.
13:20:56
ecraven
erlang might be similar, in that you can connect to a running world, modify things, and it keeps working
13:22:31
loke
ecraven: Yes, but it's more limited. You have to rebuild one module at a time. You can't just recompile a function.
13:24:16
ecraven
in a way, it is a bit dismaying to find that *no* other language can compare with CL in this aspect :-/
18:08:08
flip214
_death: but that's only a good idea for a few common sets of slots, and not 30 slots with 50% chance of existance... too many combinations
18:30:44
aeth
For me, I pretty much always use a hash table when I want a hash table. Using a(n) alist/plist for a small number of elements is implementation-specific premature optimization. However, a plist can still be a good choice with lots of elements, if you're iterating over it (e.g. alexandria:doplist)
18:33:03
_death
I never used them, but AMOP claims they're useful in knowledge representation systems
20:29:06
aindilis
what is the recommended reading path for someone who knows some lisp who wants to get good at Common Lisp for AI?
20:30:09
Bike
what kind of thing are you interested in? machine learning is the big one right now. i thought probably approximately correct was neat, but probably wrong about biology
20:30:31
aindilis
well I am interested in Knowledge Based Systems and Automated Theorem Proving, CYC etc
20:31:54
sjl
you might like Paradigms of AI Programming by Norvig (recently made freely available by him)
20:32:20
Bike
using a programming language is a good way to learn it. it sounds like you have a goal that motivates you to write code, so you're set
21:27:35
pfdietz
I've called Z3 from sbcl, but it would be nice to have a solver with more accessible innards.
21:44:59
makomo
aindilis: antoszka: i read (actually, am still reading) LoL before AMOP (haven't yet begun reading)
21:45:42
pfdietz
Applying machine learning to theorem proving is a hot topic now. https://www.floc2018.org/summit-on-machine-learning/
22:21:15
p_l
pfdietz: the summary from that link sounds more like applying formal methods to machine learning
23:48:54
White_Flame
if linux, then SBCL is great. It runs on Windows, and while it's been a bit immature there, I think it's solid at this point
23:50:02
White_Flame
there's also ECL which can interoperate in a C environment, CLASP for C++, ABCL for Java
23:58:30
edgar-rft
CLISP is newbie-friendly but somewhat semi-maintained at the moment, most people here are SBCL users
23:58:42
White_Flame
install SBCL and emacs. Then install https://www.quicklisp.org/beta/ and quicklisp-slime-helper
23:59:33
White_Flame
from there, you should be able to run M-x slime (alt-x slime on most keyboards/setups)
0:00:12
White_Flame
also, there's #clnoobs which is probably the better place for learning language stuff
0:01:46
edgar-rft
granttrec1, remove the trailing angled bracket in the headline of your browser and it probably works
0:03:17
edgar-rft
White_Flame, why not? angled brackeds are commonly used to separate links from surrounding text
0:03:50
White_Flame
irc clients have different assumptions as to what the boundaries of pasted links are
0:04:14
White_Flame
I tend to put white space around them just in case. (e.g. "Go to http://my.thing.com/foo.link .")
0:06:38
edgar-rft
that forces me to type blanks after every link, what is idiotic. I'm not responsible to fiy bugs in broken IRC clients
0:30:03
on_ion
ive been adding spaces after complete URLs for decades. web browsers like to reload from cache if next_url == current_url by string compare so it became habit to get a nice full reload.
0:35:53
edgar-rft
That's not the point. A comma *can* be part of a valid URL, an angled bracket *never* can be part of an URL. An IRC client that recognizes angled brackets as part of an URL is undoubtfully broken. Don't use shitty software.