[gambit-list] Lack of Libraries/ FFI / Tutorials

Álvaro Castro-Castilla alvaro.castro.castilla at gmail.com
Wed Dec 18 02:56:09 EST 2013

I see.
Well, clearly the second at the moment (with versioning, so you can
have different
versions in your system). These packages are going to be curated. I'm
very open to suggestions, collaborations, and participation, of
course. But I plan on using this for real projects, and I want
cohesion among the packages. That, I think, is what is always lacking
in Scheme systems. For instance, in Spheres, all FFI packages use the
same "ffi infrastructure" and idioms. I rather have 20 spheres that
work well than 150 where 30 do the same in different ways, 40 don't
work, 50 are obscure, 30 outdated...
That's why there is no release yet of SchemeSpheres, it's at
development stage. (So if you try to make it work, please bear in mind
that. I'll try to fix whatever is broken if you get errors).
Also, I believe that having SchemeSpheres designed as a "batteries
included" means that it needs to be designed, not just a bunch of

I know this has been tried many times before. The key points in this
project, that I find unique, are:
- _Curated set of libraries_, biased towards the most common
application uses (at least at the beginning).
- _Flexible & hackable_. I'm not building a Framework here. The
intention is that the hacker gets the boring things so he/she can
focus on the cool parts. But, it can be fully and easily customized.
Otherwise we are not respecting the hacker yay ;)
- R5RS plus Gambit extensions. _Only hygienic syntax-rules_.
Define-macros are only used for ver very very necessary things such as
FFI, and they are treated specially by SchemeSpheres.
- Homogeneous, cohesive, coherent. At least to some extent and as far as I

Other points:
- Grow organically, trying not to be overambitious.
- SRFIs are not in a sphere called "SRFI". They are placed somewhere
makes more sense. A SRFI that defines a datastructure is not the same
as a argument-processing one, reader extension or whatever.
- No modules until R7RS or any standard system is established. Modules
are complicated and we still don't have a good solution that works in
all cases and allows to use Gambit to its full power (No, Blackhole
doesn't allow, just try to do what I'm doing here to know by yourself,
I wasted months with that). I'm all in for a module system, but it
needs to be flexible and lightweight.
- The cool things that Gambit offers are exposed, and easy to use.
Example: remote coding/debugging in mobile apps.
- Whatever is in SchemeSpheres you can use without SchemeSpheres, just
as plain Gambit code (ie. no additional syntax in the code itself).

Of course, many of this things are work in progress.

        Álvaro Castro-Castilla    --    a at fourthbit.com
          Fourthbit: Turning Ideas into Reality.  --  Visit the Blog

On Wed, Dec 18, 2013 at 1:18 AM, Adrien Piérard
<pierarda at iro.umontreal.ca> wrote:
> Yes, portage is similar to FreeBSD's ports. Roughly:
> Ports/Portage: no every package can be updated independently by anyone
> got the rights. As a result, if things worked one month ago, and are now
> broken, you may have to revert many packages.
> Quicklisp: All packages are integrated together in a single release by the
> release manager. Tthe manager usually checked that everything builds well
> before making a release. There's much more investment there for the
> but life is easier for users for the set of packages is always consistent.
> P!
> On 17 December 2013 13:33, Álvaro Castro-Castilla
> <alvaro.castro.castilla at gmail.com> wrote:
>> I'm not a FreeBSD user, but as far as I know, Gentoo's portage is pretty
>> similar, and I'm familiar with it. I think that would be overengineering
>> the moment.
>> About Quicklisp: it seems interesting. It seems pretty similar, but I
>> don't know it in detail.
>> The tools I'm building are similar to Rake (ruby) / NPM (node.js) /
>> What I'm doing is a set of curated libraries growing organically. It's
>> everything pretty simple, though. It's better to have something simple
>> works, and build upon that. Just like Gambit does, but for the
>> developer who wants to start from something and still have all the
>> flexibility Gambit provides.
>> I'll create better documentation during the next days, so the most
>> important pieces are well documented, and can be improved on top of what
>> already done. There is a lot to be done =)
>>         Álvaro Castro-Castilla    --    a at fourthbit.com
>>           Fourthbit: Turning Ideas into Reality.  --  Visit the Blog
>> On Tue, Dec 17, 2013 at 6:21 PM, Adrien Piérard
>> <pierarda at iro.umontreal.ca> wrote:
>>> Do you plan to make something like FreeBSD's ports or rather something
>>> like quicklisp?
>>> P!
>>> On 16 December 2013 23:53, Álvaro Castro-Castilla
>>> <alvaro.castro.castilla at gmail.com> wrote:
>>> >
>>> > On Thu, Dec 12, 2013 at 7:34 PM, Răzvan Rotaru
>>> > <razvan.rotaru at gmail.com>
>>> > wrote:
>>> >>
>>> >> I wish I would have found schemespheres.org earlier. Thinks like
>>> >> should be mentioned in the wiki.
>>> >>
>>> >> Razvan
>>> >>
>>> >
>>> > Yes, that'd be nice ;)
>>> > But bear in mind that the project is still evolving and in alpha
>>> > so I
>>> > didn't talk too much about it for that reason. We are going to make it
>>> > a
>>> > solid foundation upon which you can build your own apps. But that
>>> > time
>>> > :)
>>> >
>>> > At the moment is already helpful, it should work. If you need
>>> > assistance
>>> > with its installation or you are experiencing any issues, please don't
>>> > hesitate to contact me so it can be improved.
>>> >
>>> > Best regards
>>> >
>>> > Álvaro
>>> >
>>> >
>>> >
>>> >
>>> > _______________________________________________
>>> > Gambit-list mailing list
>>> > Gambit-list at iro.umontreal.ca
>>> > https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
>>> >
>>> --
>>> Français, English, 日本語, 한국어
> --
> Français, English, 日本語, 한국어
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.iro.umontreal.ca/pipermail/gambit-list/attachments/20131218/10645d03/attachment.htm>

More information about the Gambit-list mailing list