E-guile link

[Karl Hegbloom]

On Tue, 2005-11-01 at 09:06 -0700, Tracy Brown wrote:
> 
> I have no opinion about using scheme or not for anything, but if there
> are folks who want to get rid of it; we can evaluate the use of XSLT
> as a templating engine. I've used it and it's okay but I suspect that
> eguile is much more flexible.

I kind of like the Scheme language, myself.  It is the language of
choice for introductory computer programming.  Anyone who has not
studied:

 http://en.wikipedia.org/wiki/HTDP and http://en.wikipedia.org/wiki/SICP

... is missing the bus and won't have a clue in class tomorrow.  The
opinion of the HTDP authors expressed in the foreword should be required
reading for all Computer Science and High School Mathematics educators.

I enjoy Scheme's simple syntax.  The problem with XSLT is all that < / "
= : > syntax, and the <open> </close> thing.  It's very cluttered
looking compared to the Scheme syntax.  It is harder to type due to all
of the extra characters you must remember to enter.  (Though a good
auto-completing editor mode does a lot toward alleviating that.)

If you really think about it at the level of the document structure
tree, XML and Scheme are very isomorphic.  TeXmacs has a thing where it
can show the document as a TeXmacs tree, as a Scheme file, or as an XML
file.  Theoretically, it is easily possible to write a program (maybe an
Emacs Lisp or Guile program) that reads a file written in Scheme syntax
and outputs a file in XSLT syntax.  It would be implemented with perhaps
a few reader-macros and an evaluator that outputs the XML rather than
executing program instructions.

You'd never have to worry about writing closing tags --- the closing
parenthesis would serve that purpose.  The indenting and nesting of the
Scheme editing mode in GNU Emacs coupled with the simplified, less
cluttered, syntax would make it much more readable, IMO.

If Guile is replaced in GnuCash, what will take it's place?  Parrot VM?
Certainly not a Java VM...  Dot-NET?  I'd favor the Parrot VM at this
point, I think.  That would allow, hypothetically, GnuCash programs to
be written in either Python or Perl6, and certainly a Scheme compiler
could be written for it as well if there is not already one out there.

-- 
Karl Hegbloom <hegbloom at pdx.edu>