gEDA-dev: guile-1.8 requirement

[Stuart Brorson]

My my my.... I went away on a short vacation and the list traffic
exploded!

There's been a lot of interesting discussion about guile-1.X and
dependencies.  Since all bases seem to have been covered, I'll not say
much, except,

On Mon, 4 Jun 2007, Ales Hvezda wrote:
> PS. It's still an interesting mind exercise whether or not I would use
>    guile for a new project.

Personally, I wouldn't.  I can cite two reasons:

1.  This is more minor to me, but major to others.  Guile is Scheme,
which is a minority language favored by theorists but not really well
represented in the real world.  Yes, it has some strong advantages
(like you can embedd a *small* interpreter into your program, and
Scheme is very extensible), but outside the halls of academe it is not
very common, and is regarded as arcane.  Thus, we lose many folks who
might otherwise write a back-end.  Of course, the question then
becomes: what would the back-ends be written in?

2.  This is my major gripe.  The Guile developers seem to be
contemptuous of their users.  In particular, they change the software
all the time and don't bother to support legacy stuff.  The two
problems we had  recently with guile-1.8 are representative.  In one
case, they changed the Scheme API, and in another they added a new,
spurious (for us) dependency.  In both cases, this represents a
support problem for us, the users of Guile.

Of course, Guile does have the above-mentioned advantages, and I don't
know what I'd replace it with.

I don't suggest we change now -- it's already deeply embedded into the
DNA of gEDA.  But I can still gripe!  :-)

Just MHO,

Stuart