in reply to Perl in the "Microsoft Research Center"

A lot of people (except hardburn) are missing an important point: we're talking about Microsoft Research here, not an actual product development place. The way research in company like that works is something like:

  1. "Someone" tells a researcher a topic to research.
  2. The researcher does stuff.
  3. The researcher reaches some conclusion and writes a paper about it.

There aren't any rules about which products they have to use, or whether they have to stick to the company line or anything like that. They are given a lot of freedom.

Hell, the researchers probably didn't write the actual code themselves. Many researchers just stick to the "soft" theoretical part, and leave it to a grad-student intern (a code monkey) to implement and prototype their ideas.