Re: [ANN] guile-csv 0.0.1 released!

[Nala Ginrut]

On Fri, 2013-02-08 at 08:59 -0800, Aleix Conchillo Flaqué wrote:
> On Fri, Feb 8, 2013 at 8:13 AM, Mark H Weaver <address@hidden> wrote:
> >
> > I haven't yet looked carefully at this code or its API, so this is no
> > judgement on you, but in general, I don't think we should follow the
> > model of "Hey, here's the first release of a library I just hacked up.
> > Please add it to Guildhall now."  That's how we ended up with an ice-9
> > directory that's full of bitrotted implementations of half-baked APIs.
> >
> > I'd much rather follow the example of Shiro Kawai, who is very cautious
> > to experiment with new APIs at length before adding them to Gauche, and
> > the result is IMO a beautiful and consistent set of APIs.
> >
> > Maybe we can find a good compromise position between these two extremes.
> >
> > What do other people think?
> >
> 
> I see this from the other side.
> 
> I am pretty new to Guile and I don't have much experience designing
> Guile APIs, tending to two actually (which is close to zero).
> 
> But what I see is a lack of libraries that new users can use. So, in
> my case, I just thought "hey, I want to get there but before I would
> need  a library for such and such" and I started to write them. I am
> pretty sure that people with more experience would have written a
> different API.
> 
> However, I agree that newbies should put special attention to detail,
> even more actually, to try to learn and do the best as they can. There
> should be no excuses there.
> 
> But if there are no libraries, even if not perfectly designed, it's
> going to be hard for others to use Guile. It's a vicious circle.
> 
> And if more people use Guile, they will eventually re-design some of
> the old (bad designed) APIs when they gain more experience with it. I
> guess this happens in other languages too. But, at least, there was
> something where to start.
> 

I have to choose Aleix's side this time. 
The fact is, if we don't put guile-json into the repo ASAP, "there'd be
at least two more new guile-json be written in the next year", actually
I just stop my guile-json when I see json in guildhall.
Every packages are not perfect, even if it's maintained 20 years. But if
we add the new packages without very strictly evaluate, people could
contribute based on a given start.

The bad situation could occur, if someone can't find json/csv module in
a short time, and it's not very hard to implement, they will reinvent
wheels. Or if it's very hard to implement, hey, why bother to use Guile,
let's choose Python/Ruby. Guile is interesting, but users need something
practical.

Too cautious about adding the half-baked packages would make the
language itself always be a half-baked one. In the long term, a package
will approximate to perfect if many users really need it. 

But Mark's idea is rational, so IMO the better solution is to classify
the packages, and if the submitter need to put their packages into
official quality repo, they'll face a very strictly evaluating
discussion.

What you guys think?
Thanks!

> Aleix