[FoRK] Information extraction that actually works

Ken Meltsner <meltsner at alum.mit.edu> on Wed Mar 28 18:53:33 PDT 2007

OK -- I'm an easy case because my name is almost unique (there's one
other Ken Meltsner in the world as far as I know, and only 30-40
Meltsners total).  Still, it was pretty impressive to realize that
these guys found my 3 most recent jobs, titles, etc. by grinding
through the Web.


It's been a long time, for example, since I was a member of TMS
(metallurgical professional society) but they found it.  Also my PhD
at MIT, and the positions that I held more than seven years ago.  In
fact, they only missed one professional job which was understandable
since it was nearly pre-Web.  (Left GE in spring 1993; Mosaic had just
come out)

The information was somewhat sketchy, but there wasn't any weird stuff
or junk in the fields -- missing is better than wrong in this
situation.  Given my understanding of the IE field's state of the art,
this is quite impressive.

I have no idea of how they expect to make money.  There are some ads,
there's a pro membership, and they're planning an SDK/API (it's nice
that everyone promises or provides an API these days...).  There's a
social networking aspect as well -- "colleagues" can be linked to.

In fact, perhaps that's who they should merge with: LinkedIn.  Lots of
explicit people and relationship data there, but it's handicapped by
the need to get people to update their information.

Darned impressive information extraction -- they picked up the other
Ken Meltsner as well, and as far as I know, correctly.

More information about the FoRK mailing list