Re: [Chicken-hackers] why Chicken?

[Brandon J. Van Every]

Tony Sidaway wrote:
> And being a hybrid it
> lends itself easily to the combination of low and high level
> programming that characterizes many real world applications.

That reminds me of my original reason for pursuing Scheme compilers.  
It's supposed to be important to my 3D graphics and AI problems.  The 
effort just hasn't borne any fruit yet, because I'm always building.  
I've gotten too good at building.  It's a comfort zone.

> For these reasons, Chicken isn't lacking attractiveness to excellent
> software developers.  What (apart from its relative obscurity) stops
> them using it?  It isn't the brackets, although they may drive away
> novices and intermediate programmers who have nothing to gain by the
> effort of learning a new language unless it provides them with access
> to a huge wealth of library software that would otherwise be out of
> reach.  No, Chicken is hard work.  It's pitched at the Lowest Common
> Denominator, Posix, and if you want to do anything useful the first
> thing you must do is produce a set of language bindings.  While this
> isn't too difficult in Chicken, it is time consuming work and it's the
> sort of thing developers expect to come as part of the package with
> any modern language.
>
> There should be a set of bindings for various native OS features.
> Staying in the Posix ghetto was an attractive option ten years ago
> when the most widely used version of Windows was little more than a
> toy, but if Chicken is to be useful for producing widely-used software
> that exploits the full capability of modern systems it must provide
> the kinds of interfaces that developers need, and that means more
> bindings to native features, in particular native Windows features and
> CLI (.Net) features.

That would be Bigloo Scheme.  It has C, Java, and C# back ends.  And as 
far as I'm aware, Bigloo is even less popular than Chicken.  So how is 
this really an answer, to any strategic question of Chicken's future?

I think problems of popularity cannot be solves with tech.  They are 
solved with marketing.  I've been through language marketing efforts 
before, and failed at them.  I don't feel ready to burn a lot of time on 
that problem.  CMake still isn't rock solid on Linux, the docs don't say 
the right things about CMake vs. Autoconf, and I still haven't done 
anything significant with Chicken to prove its utility. 

>   If these were provided in eggs the clean
> simplicity of the core language implementation would not be affected.

If there's a person who wants to do it, that sees it as their mission, 
then things could happen.  Otherwise, it's wishful thinking.

I would comment that Chicken is unlikely to be better at talking to Java 
than Kawa Scheme is.  I don't see any competitive advantage arising there.

Not many Schemes bother with C#, so I see more possibility of market 
differentiation in that.  Let's assume that adequate C-to-C# interfaces 
are possible, to get real work done.  There's still a strategic problem: 
C# is still a Microsoft trap.  Although the language itself is an ISO 
standard, nobody's interested in C#, they're interested in .NET.  That's 
wholly proprietary and Microsoft can yank out the rug at any time.  So 
you need people ideologically committed to .NET to form the basis of an 
open source effort.  You need Windows developers.  And I think we have 
precious few of those around here.  The reason I just spent a year 
banging on the Windows builds, is no one else had come along to do it.  
Now, the Chicken and eggs could get better.  With CMake, I've laid the 
groundwork for more Windows usage.  But I don't see the C# .NET problem 
as being a quick and easy implementation.  I think more Windows 
developers first have to be attracted to Chicken for other reasons.

On a personal note, the .NET CLR takes people in directions I don't 
personally care about.  If I had wanted to do CLR stuff, I would have 
done it 3 years ago.  But the CLR is not performance oriented, and leads 
to things like web apps, accounting software, business apps, etc.  Just 
not my cup of tea.  I'm a 3D / AI / game developer.  C# has yet to prove 
any industrial relevance in those arenas.  I think Managed DirectX is a 
joke.  I'm willing to be surprised, it's been awhile since I've bothered 
with the C# gaming universe.  But Microsoft would have to show me an 
awful lot, before I'd remotely care.


> On the UNIX side, a similar lack of Gtk/Gnome bindings and the like
> makes Chicken programming quite hard work.  Similarly they need not
> concern Chicken's core.

Is there a Gtk/Gnome person around here that wants to worry about it?  
Otherwise, again, it's wishful thinking.

"Let's do more binding!" does not solve the strategic problem of ongoing 
gruntwork.  In fact it does the opposite, it creates tons of it.  We 
need something that solves / alleviates gruntwork problems.  I see two 
approaches:

- a recruitment strategy that attracts massively more open source 
developers to Chicken.  So that we get out of the tens and into the 
hundreds of developers.  If I had an entire team working on build 
issues, we wouldn't have any build problems, and nobody would be doing a 
ton of boring build work.

- a brilliant technology that automates binding work.


Cheers,
Brandon Van Every