Re: Roadmap and goals?

[bitwize@wizards-of-source.org]

On 17 Apr 2002, Neil Jerram wrote:

> 
> Some of the things that interest me are: documentation, both manual
> and online; debugging facilities; Emacs integration; libguile
> factorization; Elisp translation.

It may be too much to ask, but I'd like to see a VM being integrated into 
guile-core. The one that Keisuke Nishida wrote a while back is supposedly 
really good; integrating it so that byte-compilation and execution of 
bytecodes becomes the default method of interpretation and execution (much 
as it is with mzscheme) would be a Good Thing for Guile, performance-wise.

Another thing is, I use Guile where many hackers would choose Perl or 
Python, as an all-purpose automation and glue language. It's there, it 
works, and Scheme is much easier to write things quickly in than those 
other languages. Anything that boosts Guile from a systems integration 
standpoint would be nice for my uses. Responding to .NET, SOAP, and all 
that hype-based hoopla with efficient Scheme-based simple RPC and 
secure mobile code mechanisms would be nifty.

That's my wish list for the time being. The TODO also mentions a decent 
Guile binding for Tk. I'm currently working on this using pTk; has anyone 
gotten any further?

> For me I think the answers are (i) because it's GNU (ii) because it's
> what I'm now familiar with, which makes me inclined to stick at it
> until we get it right.  Plus in all the things that I've tried to do
> with it, I've not yet hit any insurmountable technical or
> philosophical barrier - it mostly "just works."

The fact that "it's GNU" is perhaps instrumental to Guile's success. 
Scheme has overall been slapped with the Pascal stigma: it's an 
"educational language", which means it lacks utility in the Real World. 
Guile represents disproof by counterexample, by being quite central to 
many of the things GNU is doing (GNOME, etc.).

> Allegedly Guile is good for integration with C projects, but to be
> honest I haven't evaluated the competition.

SIOD may actually be better in this regard, because of its tiny size and 
liberal license. But then again, SIOD lacks RnRS compliance, and much of 
the other niftiness that Guile has.

-- 
Jeffrey T. Read
"LOBSTER STICKS TO MAGNET!"


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel