Re: Considering Guile as part of a Master's Thesis

[Pat Eyler <peyler@nortom.com>]

Bradley,
I'm not sure what you'll see from the rest of the community, but the
following are some things I think are important to Guile and not there
yet.

   * multiple syntaxes -- one of the early promises of guile was that
     you could write in multiple languages (TCL, Perl 5, etc.) and run
     it in the Guile interpreter [see note 1 below].  Implementing one
     of these (either BASH [see note 2 below] or Perl 5) would be very
     cool [see note 3 below].
   * a byte compiler -- this has been talked about on the list (as has
     a true guile compiler or a guile to C compiler), check the
     archives at www.red-bean.com/guile/guile for more info.
   * better tools for handling specific things -- there are a few
     packages available on the ftp archive in the contrib area, but
     more would be nice, for example:
	  - snmp
	  - html parsing
	  - translator development tools [see note 1 below (again)]
   * reriting dejagnu to use guile [see note 4 below]

-pate
your friendly neighborhood arachnist


[note 1] there is an example syntax (ctax) and some tools for this
that were a part of early guile.  I'm not sure if/how well these work
anymore. 

[note 2] rms wants to have guile built into bash.  Making guile
understand the bash sytax and making it capable of handling multiple
syntaxes concurrently would go a long way toward fulfilling this
desire.

[note 3] making guile understand expect would be very good as well,
some work in this area was done in pre 1.0 guile.  I don't know if
this stuff still works.

[note 4] if guile were to understand expect syntax (a la note 3), this
would come out in the wash - wouldn't it?