Re: a passionate guy who want to join in as a developer

[Mark H Weaver]

Hi Rushan,

On 08/06/2012 10:50 AM, rushan chen wrote:
> I'm just a guy who is passionate about programming especially
> programming in C, and want to enjoy it with people like-minded. Guile is
> an interesting project to me, and I really really hope I can make some
> contribution to it as a developer.

Sounds great, we can certainly use more help :)

> Hope someone could point me some direction, thanks in advance.

It's hard to make specific suggestions without knowing more about your 
skills and interests.  Most Guile developers started out by using Guile 
for their own work, and ended up fixing bugs that they found, and added 
features that they needed.  I think this is the best path, because it 
makes it far more likely that you will understand the needs of users who 
need a particular feature.

There's been a recent trend in Guile to write more code in Scheme and 
less in C.  Most of the new C code is fairly deep in the guts of Guile, 
and I'm reluctant to point you in that direction until you've gained 
some experience working on Guile.

If you prefer to write C code, then one category of useful jobs is to 
write bindings for C libraries so that they can be easily accessed from 
Scheme.  Of course they should be C libraries that Guile programmers 
would like to use, and that would not be better to implement directly in 
Scheme.

Another thing we need that makes sense to write in C is a larger library 
of efficient data structures.  For example, we could really use an 
efficient balanced tree implementation.  However, if you want to work on 
a new data structure, please first post a proposed API to this list so 
that we provide feedback before you get started on the implementation, 
to avoid wasted work.

Here's one specific data structure idea: we have a "vlist" data 
structure, currently written in Scheme, but in my opinion it would be 
useful to rewrite it in C for better efficiency.  It is already used 
quite heavily in our compiler, so you could make our compiler run much 
faster this way.  Also, users might be more inclined to use vlists if 
they were more efficient.

And of course, I should invite you to take a look at our bug tracker
http://debbugs.gnu.org/guile and see if there are any that you can fix. 
 While this work is not as fun as adding new features, it is deeply 
appreciated and a great way to learn about Guile internals.

Well, those are a few ideas anyway, but I still think the best path is 
for you to scratch your own itches.  Use Guile for your own work, report 
bugs that you find (and fix them if you can), and more generally try to 
find ways to improve Guile that would have made your work easier and 
more fun.

Thanks so much for your interest, and happy hacking!

     Mark