I'm a little confused. Do you want unicode characters to be considered
letters to be parts of names or do you want to treat them as primitives?
Take xxøabc+de . How should the ø be treated?

On Sat, Feb 12, 2011 at 5:18 PM, Dan Bron <[email protected]> wrote:

> John asked:
> >  What kind of word-formatting
> >  do you want to do?
>
> Sorry, I meant word-forming, as in  ;:  .  My initial thought is to
> (minimally) modify the FSM given in the definition of monad ;: such that
> sequences of (appropriate) Unicode characters are recognized as words.
>  I'll
> then Punycode-encode those words, and pass the resulting (pure ASCII)
> sentence on to J for execution.
>
> In the next iteration, I'm going to replace certain J primitives and
> standard functions in the input sentence with cover functions containing
> Unicode-preprocessors, to allow the expected level of meta-programming (for
> example, I'd need cover functions for  ".  and  load  ).  I'm also
> considering the wisdom of applying these substitutions to the entire
> existing namespace, when my framework is loaded.
>
> Though the approach has its limitations and blind spots, it seems to me to
> be the easiest way to implement a J/APL-like language on top of J (i.e.
> that
> is, I'm trying to avoid writing a full parser by reusing J's tools).
>
> -Dan
>
>
>
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to