cl-autowrap VS paip-lisp

> Lack of access to the C libraries.

???

I recently started learning Common Lisp for fun (and fun it is!) and the ease of accessing C libraries was one of the things that surprised me in a positive way.

Using https://github.com/rpav/cl-autowrap one can simply write (c-include "file.h") and the API defined in "file.h" is accessible from Lisp. I can't think of a simpler way.

Even without cl-autowrap, FFI using https://cffi.common-lisp.dev/ seems simple enough.

I think the closest is cl-autowrap. I can imagine a higher level wrapper around it by which it can translate the python header file into the CL counterpart, although I'm not sure how much work the translation might entail. Also, because python and lisp semantics can differ considerably, the generated code might be trying to do weird things - again an issue of translation.

Common lisp has a "pretty OK" story for calling C code whenever some speed is needed [0,1]. In my opinion, they suffer from some of the documentation/quick start problems that common lisp has, but they're otherwise usable.

Some of Naughty Dog's late 90's/early 2000's games (Jak and Daxter, Jak II) were written in a lisp called GOAL, Game Oriented Assembly Lisp [2]

[0] https://github.com/rpav/cl-autowrap

The closest thing to what you request, that I'm aware of, is cl-autowrap (to use C code from Lisp) but it is not standard in any way. CFFI is the de facto standard for using C from Lisp across different implementations.

:claw tracks back to 2017 as a fork of cl-autowrap with cl-autowrap/pull/83 feature.

If you're interested in FFI, then yeah CFFI is the standard. The other comments addressed speed, I also wanted to point out https://github.com/rpav/cl-autowrap which is built on top of CFFI and can help get a wrapper up and running faster. After using autowrap's c-include you can then use CFFI basically like normal or some useful autowrap/plus-c's helper functions -- e.g. in one project, I have an SDL_Event (https://wiki.libsdl.org/SDL_Event) and to access event.key.keysym.scancode I have a helper function that's just (plus-c:c-ref event sdl2-ffi:sdl-event :key :keysym :scancode). Last year I wanted to try out using FMOD, and even though it's closed source and has a (to me) "interesting" API things worked easily: https://gist.github.com/Jach/dc2ec7b9402d0ec5836a935384cacdc... More work would be needed to make a nice wrapper, type things more fully, etc. but depending on the C library you might find someone's already done that (or made a start) and made it available from quicklisp.

In recent years there has also been cl-autowrap; caveats -

There is the cl-autowrap that can generate lisp packages from C header filesc- I am unsure if it sticks to ANSI C or goes beyond. It inturn depends on c2ffi for the first time around.

Have you seen https://stevelosh.com/blog/2018/08/a-road-to-common-lisp/ ? "Kludges" everywhere is applicable. On the other hand, having a function like "row-major-aref" that allows accessing any multi-dimensional array as if it were one dimensional is "sweeter than the honeycomb".

I still think CL code can be beautiful. Norvig's in PAIP https://github.com/norvig/paip-lisp is nice.

As for the inside-out remark, while technically you do it, you don't have to, and it's very convenient to not do. Clojure has its semi-famous arrow macro that lets you write things in a more sequential style, it exists in CL too, and there's always the venerable let* binding. e.g. 3 options:

    (loop (print (eval (read))))

PAIP by Peter Norvig, Chapter 23, Compiling Lisp

https://github.com/norvig/paip-lisp/blob/main/docs/chapter23...

Emacs is so much more than a text editor! But I need to stay on topic...

I believe your assessment of LISP (and therefore of MacArthy)'s impact on AI to be unfair. Just a few days ago https://github.com/norvig/paip-lisp was discussed on this site, for example.

Sounds like a great project idea to make a toy demo of this direction you'd like to see. Maybe comparable to https://github.com/norvig/paip-lisp/blob/main/docs/chapter15... and https://github.com/norvig/paip-lisp/blob/main/docs/chapter8.... which are a few hundred lines of Lisp each, but do enough to be interesting.

You could look into Paradigms of AI Programming by Peter Norvig which might interest you regardless of Lisp content.

Lisp has great power, examine PAIP, part II chapters 7 and 8.

A computer running Prolog reasons, and that only requires a couple of pages of code. So it seems feasible that the network could have learned some ability to reason within its network.

IMHO it's because lisp shines to manipulate symbols whereas the current AI trend is crunching matrices.

When AI was about building grammars, trees, developing expert systems builds rules etc. symbol manipulation was king. Look at PAIP for some examples: https://github.com/norvig/paip-lisp

This paradigm has changed.

Origen: Conversación con Bing, 4/4/2023(1) gigamonkey/monkeylib-binary-data - GitHub. https://github.com/gigamonkey/monkeylib-binary-data Con acceso 4/4/2023. (2) paip-lisp/chapter4.md at main · norvig/paip-lisp · GitHub. https://github.com/norvig/paip-lisp/blob/main/docs/chapter4.md Con acceso 4/4/2023. (3) bibliography.md · GitHub. https://gist.github.com/gigamonkey/6151820 Con acceso 4/4/2023.