[developers] divergences between `old' and `new' LKB branches

Ann Copestake Ann.Copestake at cl.cam.ac.uk
Mon Jun 20 16:23:29 CEST 2005


ok, but I may not be able to look at it immediately

packing - yes I thought about that - see note in parse.lsp - but wanted to get
it working first.  Obviously that's one reason for moving to a chart all the
way through.

I discovered one problem with the code last night, which I'll check in a fix
for when I get a chance.  This was potentially preventing some valid 
analyses of a string in the situation where one rule was fed by some rules and 
another rule associated with the same affix was fed by others.

I am trying to merge in irregulars which is complicated by the need to allow
irregulars where the derived form is equal to the stem.  My current feeling is
that this should only be allowed in irregulars - if there's a regular pattern
where there's no affixation, this has to be allowed for by a lexical rule.

Ann




More information about the developers mailing list