freenode/#lisp - IRC Chatlog
Search
16:54:18
theseb
You can really define the entire lang in terms of about 8 primitives? Why does that seem impossible to me? Think of how complex common lisp is
16:58:16
Shinmera
The semantics, yes. But an implementation still needs to do a lot more -- like handling OPEN, I/O, etc.
16:59:03
theseb
Shinmera: aha! yes..i knew something was missing!!! All the //side effects// were not mentioned like I/O!
16:59:55
MichaelRaskin
Well, if you target Plan9 you don't need many I/O primitives, and then they cover all the OS interaction…
17:00:03
pjb
theseb: with lambda calculus, you can implement state and I/O by simulating the whole universe!
17:02:38
pjb
theseb: there's a "desire", or "universal love" or "god" that makes kinds of quantum fluctuations in the mathematical world generate random mathemathical objects. Functions and lambda calculus are actually very simple mathematical objects, and thus appear naturally, and "develop" (are "created") by themselves, until find this god lambda calculus program that decides to simulate the whole universe and to reproduce little simulatio
17:09:34
pjb
theseb: also, it's not very hard to implement a hardware simulation to do I/O in pure lambda calculus.
17:11:50
theseb
pjb: imho..i think it best to just punt and define an abstract "input" and "output" primitive
17:12:40
pjb
theseb: of course, if you want to hook to other universes, such as ours, you can always branch out to primitives implemented in another system, eg. in electrons and silicium atoms.
17:13:28
pjb
(but notice the funny nature of fermions and bosons that are both just probability wave functions, ie. not any more real than lambda calculus functions).
17:13:38
theseb
pjb: what the #$@#$? why are you making this so complicated? i'm suggesting a simple input and output primitive while you want to hook up to other universes!?!!? am i missing something?
17:16:25
pjb
Primives, like axioms, are your own totally and entirely subjective and arbitrary choice.
17:18:42
theseb
pjb: i like that link....avoid english and use lispy mini-langs instead to bootstrap
17:19:36
pjb
or realize that lisp was created 50 years ago, and that we don't need to bootstrap anymore: just write CL in CL.
17:20:09
pjb
You might need to bootstrap again if you go to Mars, and a meteor destroys all computing equipment.
17:20:33
pjb
Then you will have to mine metals, build a new computer, and program it, bootstrapping, hopefully, a new lisp system.
17:20:58
theseb
pjb: actually...imagine we worked in a sector where security was paramount.....i wonder if defining layers of langs somehow avoids potential security holes
18:27:45
drmeister
Within clasp I am going (require :asdf) --> loads clasps shiped 3.3.1.2 version of ASDF
18:29:08
drmeister
Why is it telling me I have an older one registered at sys:modules;asdf;asdf.asd ?
18:36:45
pjb
drmeister: it helps testing for a version, or higher, since in general releases are backward compatible.
18:36:55
drmeister
Probably not - I don't know what all those features are for - they may not be related to the warning.
18:37:17
pjb
drmeister: ie. ASDF 4.0 will probably be compatible with asdf 3.3, asdf 3 and asdf 2, so all those features will have to be present.
18:38:13
drmeister
Do you know why the warning is being generated? I cloned quicklisp-client.git from github and I'm using that. When I load the quicklisp:setup.lisp I get that warning even after I've loaded an asdf.
18:51:01
jackdaniel
it is a logical pathname to something like /usr/lib/clasp-x.x.x/modules/asdf2.fasl
18:51:56
jackdaniel
and warning is being generated from ASDF itself, it doesn't like its previous incarnations. it is not about asdf2 being loaded, but mere about being present in the path
21:23:12
drmeister
I'm dropping source code and a cache directory full of compiled files into a docker image and then running code.
21:23:38
drmeister
asdf is rebuilding everything even though the timestamps are preserved - so I'm a bit puzzled.
21:47:16
drmeister
Is there a function like (asdf:why-the f*ck-are-you-rebuilding-everything system)?
21:49:09
Shinmera
You can manually build an ASDF action plan (I showed you how once before) -- the plan should only include things that need to be done, excluding already performed operations
21:49:24
Shinmera
Then you can try tracing select functions in ASDF to see where it's getting things from
22:04:53
AeroNotix
is there an asdf function that can find the modules that the `:components` key describes. Alternatively, a function that gives me the asdf definition?
22:07:23
Shinmera
If you want the list of components in the system definition, asdf:component-children
23:04:59
aeth
dim: Idk, I think the most graceful way for an IRC client to fail is to pretend like it's working, complete with fake conversations. The conversations are even predictable: talk about how the IRC client is now working.
0:03:33
dwrngr`
I think of it like the street slang of programming. It has a consistency to it and you have to keep up somewhat in order to fit in. But it gets in the way in a professional setting :P
0:06:19
dwrngr`
I guess Typescript is supposed to make it more tolerable there, but personally the limited times I have to use it I've found parenscript to make life a lot simpler
0:44:37
aeth
(And if you wanted to do a TypeScript-like-thing in CL all you'd need is a handful of macros)
2:27:54
kenster
doing some live lisp programming here: https://youtu.be/l4Uj13QLtpU or https://www.twitch.tv/kingherring
3:07:11
siraben
You might see "Quit" in the minibuffer time to time when he presses C-g to cancel selection
3:12:57
mason
If you *really* want to achieve flow, though: https://www.youtube.com/watch?v=sBvt0lsoInc