Google Programming Contest

[Grant Barrett]

I'd like to bring this list's attention to the Google Programming Contest,
in which people are asked to write code for new ways of processing the
Google index. There's a $10,000 prize.

http://www.google.com/programming-contest/

Now, at first glance this doesn't seem appropriate here, but from my
perspective, I see the Google index is the first resort for a lot of the
word-related work going on. In addition, I suspect there are more than a few
experts of computational linguistics among us who may have the skills to put
the Google index to a lexicographical or linguistic use. It seems like a
great opportunity.

Some ideas, freely given, perhaps worth combining:

-- A Google search that returns results for specific word or phrase in
typical concordance fashion. The results would be lines of linked text,
taken from the original document, and centered on the word or phrase
searched for so that they line up down the middle. None of that other
information now returned would appear in order to free the results of
clutter.

-- Exact phrase searches which are punctuation- and case-sensitive

-- Soundex searches

-- An estimated quality meter, in which certain rules for grammar, spelling,
punctuation, clichés and rhetoric are applied against a search result,
ranking it according to the absence of these red-flag items. The list of
red-flag items is endless, but a good core of 500 or so rules shouldn't be
difficult to create and could be useful.

Anyone?

PS: This contest is being discussed at length among the geek set:
http://slashdot.org/article.pl?sid=02/02/06/2025229&mode=flat