On Wed, Nov 14, 2007 at 11:43:52AM -0500, Brion Vibber wrote:
In other words: let's record the syntax as it exists in people's minds (and their existing work), rather than the behavior of the actual parser.
Bingo. :)
And the suggested approach for actually doing that which is *not* "grind over the current corpus and see what constructs are actually being used[1]"... is?
Cheers, -- jra [1] as I suggested 327 posts ago :-)