language design tradeoffs and tcl

I received many responses from my post requesting some help with language
design tradeoffs. Most of the responses fell into two categories : use
some sort of command completion to speed up command line input (this is
the solution I'm going with); and use tcl.


The problem with tcl is twofold. First of all, its only a scripting
language. It doesn't provide good parameter passing mechanisms or
flexible data structure types--everything is a string. The way I phrased
my question, however, certainly made tcl a reasonable approach (one that
I'll probably use in the future).


The main problem with tcl, though, is that everthing is a string, and not
just a string, a char *. This is not amenable to achieving
internationalizable software, a primary goal of the project I'm working
on. This is an issue that receives very little consideration in current
languages, which usually just leave it up to library routines. Over time,
I18N will become a more and more important goal for all software however.


Thanks for all your responses!
--
jeff
kotula@cs.umn.edu
--

Return to the comp.compilers page.
Search the comp.compilers archives again.