Brian Behlendorf <brian@collab.net> writes:
> On 2 Feb 2001, Jim Blandy wrote:
> > pohl <pohl@screaming.org> writes:
> > > > I thought this was pretty interesting.
> > > >
> > > > http://people.redhat.com/~hp/proposal/
> > >
> > > Wasn't this of the original vision for guile?
> >
> > Yes, kind of.
>
> I think every programming language of note has at one point or another
> entertained notions of being a universal "glue" for everyone else. The
> problem is, a universal glue presumes a programming model and/or syntax
> and/or widespread implementation and/or license that everyone finds
> usable. Unfortunately, between rooting for the underdog and resisting
> change, no single glue language/layer has ever come to pass. But I think
> Havoc's got a grip on problems, at least, which is more than most who have
> passed this way.
Yep.
In RMS's defense, he's not actually much of a language bigot. Guile
was supposed to provide "translators" for other languages, effectively
turning the Scheme engine into a VM for anything you wanted. This is
similar to the .NET design, which includes a GC and VM. But nobody's
gotten any translators plugged in yet. I think it's mostly a lack of
technical leadership and initiative. Guile has nobody with the
vision, energy, and time (plenty of people with one or two of each,
though) to give it real impact.
What I think gives Microsoft the advantage here is that they have the
centralized power to declare, "*This* is the common standard to which
you shall program" whereas Open Source has always pretty much said,
"Cats don't need herding."
Received on Sat Oct 21 14:36:21 2006