[developers] support for default unification / YADU

[John Carroll]

Hi,

I'm wondering whether it's worth preserving support for default feature structures in the LKB. Has anyone tried to use defaults in the recent past - and is it likely that anyone will want to do so in the future? Since I've started the LKB-FOS effort I've tried to retain this facility, but I've never been able to verify that it actually works since I've never tested grammars containing defaults.

I'm asking about this now, since there's a change I want to make to the LKB that will probably irretrievably break default unification in both parsing and generation.

Any opinions? Please reply if you feel strongly one way or the other.

Thanks, John