Spinning Yarns

I could not pass up relating the following amusing anecdote from this afternoon.

After the faculty candidate had finished giving his talk, there was a question and answer period. One of the interesting applications of his research is to figure out the shape of an object based on points sampled from within the object’s volume. To do this, the basic idea is that you construct a spherical region around each point, and build an approximation to the original object by figuring out where these spheres intersect. In topological terms, such a sphere is called an epsilon-ball, since it has radius epsilon (ε). The quality of the approximation you get depends, in large part, on the value of epsilon you choose.*

* This was not, in itself, the really interesting part of his work, just a basic tool for the problem he was talking about.

The candidate had mentioned that figuring out what radius to use is a difficult question in its own right, and so during the Q&A, one of the audience members raised his hand and asked, deadpan, the following question:

You mentioned earlier in your talk that there is a question of how to choose epsilon. If you have large enough balls, won’t virtually any object wind up looking like a big sphere?

Now, maybe it’s just me, but that was the best question I’ve ever heard anyone ask a faculty candidate. Maybe it should become a standard test: “Are your balls large enough for this job?”

Or maybe, just maybe, I should grow up. Nah.

Followup, June 2007. It turns out that this candidate was given an offer, and, after deferring for a year or so, showed up on campus this past Fall as one of our new faculty members. So, apparently, the balls were big enough, or at least impressive enough.

Filed by Michael at 19:25 under Miscellaneous
1 Comment