I've talked to a couple of you about this at the Barcelona summit but I wanted to send a broader email.<div><br></div><div>One thing I'd love to have is a fast, efficient, library-based implementation of feature structure unification. Something written in C and separate from the cheap code so that it could be called by other programs/scripting languages, NLTK etc.</div>
<div><br></div><div>I'm picturing a user-provided function for value unification whose prototype looks something like this:</div><div><br></div><blockquote class="webkit-indent-blockquote" style="margin: 0 0 0 40px; border: none; padding: 0px;">
bool value_unify(char *value1, char* value2)</blockquote><div><br></div><div>so that the whole issue of what kind of grammar you have is kept separate.</div><div><br></div><div>Has anyone ever tried to create a library like this? If not, where is the unification source code in cheap so that I can get an idea of how it works?</div>
<div><br></div><div>I have absolutely no time to look at this in the near future, but I've been curious about it for a while so it would be good information to know and have archived.<br clear="all"><br>-- <br>W.P. McNeill<br>
<a href="http://staff.washington.edu/billmcn/index.shtml">http://staff.washington.edu/billmcn/index.shtml</a><br>
</div>