[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: good papers on graphics in scheme/lisp?

From: Blake Shaw
Subject: Re: good papers on graphics in scheme/lisp?
Date: Fri, 04 Mar 2022 02:10:15 +0700

Olivier Dion <> writes:

> On Wed, 02 Mar 2022, Blake Shaw <> wrote:
>> Next week I will start working on my installation /Scribble Jam/ for
>> iLight Singapore[2], and I'm planning to do it all in Guile OpenGL +
>> GLSL. I've started fiddling with Guile OpenGL and it seems pretty
>> complete and performant (I haven't experienced any GC hiccups from the
>> small stress tests I've put it through, but its still only minor
>> stuff).
>> There is still a chance I might choose to create a Guile
>> interface to a subset of the C++ OpenFrameworks[3] library
> FFI for C++ is a real pain to do.  You can't juste use plain Scheme (at
> least I don't know how for C++).  If you go that route, I have C++
> templates that can help you for the glue between Scheme and C++
> (e.g. std::vector, std::map, std::string).

Thanks for this heads up, I've been curious about the C++ FFI experience
in Guile (while less excited about actually getting into the
weeds). If I go this route, I'll probably wind up approaching it from
using libguile, adding a guile interpreter as an abstraction layer over 
the creative coding I'm doing in C++, as my experience with that has so
far been pretty straight forward (in pure C). But I also worry that if
I go down this route, out of habit and familiarity I'll for the most
part just end up doing the whole thing in C++, with Guile acting as more
of a sort of controller for livecoding. Which could end up really great!
I'll be putting a couple days aside this coming week to take it for a
test drive with old projects to see how it feels.

But after some hacking tonight I'm feeling pretty excited about the
pure-Scheme route, so if it appears feasible after weighing my options
it will certainly be the direction I head in.
>> Also, one last thing, I'm considering attempting to create a DSL that
>> compiles to a subset of GLSL like Chicken's GLLS library[4], as I'm
>> becoming more and more interested in compilation. I remember reading
>> somewhere about a Guile library that allows you to write C in Guile,
> I think you are refering to
> <>.
Yes! thank you so much, this is it. And looking through the code base I
can see there are already some pieces there that seem to illustrate how
to work out formatting issues I was confused by as a newcomer to
srfi-159 (which is amazing, I might add for folks searching the
archive. go check it out ASAP:). I've also come across srfi-166, which
is Alex Shinn (same author)'s monadic formatting procedure that deprecates
fmt/srfi-159 in order to improve accuracy and safety, which apparently
caused some minor headaches in glls. so I'll probably be porting
srfi-166 to guile as well :)

I've actually gone ahead and started porting GLLS to Guile and to my
astonishment its been a piece of cake so far. If I keep up this pace I
could even have a working prototyp by the end of the weekend -- but that
just seems to good to be true (am I finally acquiring the famed lisp
super powers? tbd) 
> There's also <> that can be helpful if
> you're familiar with LALR(1) parsers.
> Regards,
> old
yes! nyacc has been extremely helpful. great software, my thanks to the
authors, I was even able to use it to wrap glValidateProgram, which
seems to be fine "out of the box", so to speek.


“In girum imus nocte et consumimur igni”

reply via email to

[Prev in Thread] Current Thread [Next in Thread]