Probably because there's no easily accessible data store like IMDB for famous mathematicians or computer scientists.
http://www.ams.org/mathscinet/collaborationDistance.html
using the MathSciNet reference database. Unfortuately, it is only accessible to MathSciNet subscribers, so basically on all university campuses or libraries.
My Erdos number is no higher than 5 :-).
http://academic.research.microsoft.com/VisualExplorer
(The gimmick is that Erdos is the default for the co-author path visualization)
http://www.freebase.com/view/en/computer_scientist/-/people/...
The more important piece thats missing is the one that Journals need to contribute - papers and author lists. ;)
It would be surprising to see any suit to enforce the licensing restrictions on a database built from user contributions like IMDB actually go to trial.
We might come away with some interesting answers to the questions of what licensing terms on such databases are truly enforceable (as opposed to mere intimidation).
What happened to that Craigslist lawsuit? How is that progressing?