April 14, 2002:

I haven't worked on Yapps for a while, mainly because I spent all my energy
on trying to graduate.  Now that I've finished school, I have several projects
I want to start working on again, including Yapps.

Notes for myself:

Add a debugging mode that helps you understand how the grammar
  is constructed and how things are being parsed
Look into an English output mode that would use natural language
  to describe a grammar
Optimize unused variables
Add a convenience to automatically gather up the values returned 
  from subpatterns, put them into a list, and return them
Improve the documentation
Write some larger examples
Get rid of old-style regex support
Use SRE's lex support to speed up lexing (this may be hard given that
  yapps allows for context-sensitive lexers)
Look over Dan Connoly's experience with Yapps (bugs, frustrations, etc.)
  and see what improvements could be made
Add something to pretty-print the grammar (without the actions)
Maybe conditionals?  Follow this rule only if <condition> holds.
  But this would be useful mainly when multiple rules match, and we
  want the first matching rule.  The conditional would mean we skip to
  the next rule.  Maybe this is part of the attribute grammar system,
  where rule X<0> can be specified separately from X<N>.
Convenience functions that could build return values for all rules
  without specifying the code for each rule individually
Patterns (abstractions over rules) -- for example, comma separated values
  have a certain rule pattern that gets replicated all over the place
"Gather" mode that simply outputs the return values for certain nodes.
  For example, if you just want all expressions, you could ask yapps
  to gather the results of the 'expr' rule into a list.  This would 
  ignore all the higher level structure.
Look at everyone's Yapps grammars, and come up with larger examples
  http://www.w3.org/2000/10/swap/SemEnglish.g
  http://www.w3.org/2000/10/swap/kifExpr.g
  http://www.w3.org/2000/10/swap/rdfn3.g
It would be nice if you could feed text into Yapps (push model) instead
  of Yapps reading text out of a string (pull model).  However, I think
  that would make the resulting parser code mostly unreadable 
  (like yacc, etc.).  Coroutines/stacklesspython may be the answer.