Tue Mar 29 09:17:38 EST 2005
> Since my October 19 note, where I laid out some of the difficulties,
> I've done some research on existing general-purpose matchers and thought
> really hard about how we might make this work. While I still agree that a
> unified matcher is a desirable goal in principle, I have become convinced
> that it's highly unlikely that we can come up with a general-purpose
> matcher that also doubles adequately as a domain-specific matcher for
> syntactic abstraction. At best, we could specify a straightforward
> translation from something like syntax-case into a general-purpose
> matcher, assuming some level of similarity. Furthermore, given the
> variety of matchers out there, I suspect we'd even have trouble agreeing
> on a general-purpose matcher independently of syntactic abstraction.
Why do you consider that its hard to use a unified matched? Could you give
me some hints? Is it because of the problem of keeping tracks of the location
in the source files of the original s-expressions?
More information about the R6RS