gEDA-dev: guile-1.8 requirement

[Svenn Are Bjerkem]

On 6/4/07, Stuart Brorson <sdb at cloud9.net> wrote:
>
> I like Python plenty.  The requirements for a Guile replacement are:
>
> 1.  Can easily embedd small interpreter into gnetlist.
> 2.  Language is adapted to handling netlists (i.e. deals with text
>      well, and one can create reasonable  data structures with it).
> 3.  Language is well-known in EDA community.
> 4.  Developers/sponsors of language understand and are sensitive to
>      the needs of their users.
>
> Of the above, only #1 and 3 are questionable to me.  Yes, you can
> embedd a Python interpreter, but is it small?  (Note that linking to
> Python doesn't make me happy since it then re-introduces the
> dependency problem.  Moreover, Python does change from release to
> release, although it's not as bad as Guile.)
>
> As for #3, I don't know if the folks who use gEDA (as opposed to
> software developers) are up to speed on Python.  I do know that TCL is
> popular out there with EDA types.  Also, TCL is kinda like Lisp in
> that it can behave as a functional language.  However, it isn't
> strictly functional since you can also write procedures with it.
>
> Anyway, I don't think we should change, but thinking about
> alternatives is a fun exercise.  And folks from other projects
> Googling for info on embeddeding an interpreter into their software
> might benefit from our experience via reading this thread.

Ales asked for what to use in new projects. And if he ever put up an
email voting, I would not vote for guile. I cannot free my mind from
the fact that Guile is a protest against Tcl and as such will not have
support from the EDA community.

Tim Edwards, who does work on xcircuit and magic, uses Tcl as glue
language after having tried Scheme and Python as embedded languages.

-- 
Svenn