I was recently listening to a FLOSS Weekly podcast (episode 100, congratulations!) interviewing Chris DiDona from Google. One of the items he spoke of that resonated with me was his/their practice of deciding who they wanted to reach and focusing on that audience even if it meant delaying or forgoing contact with others. DiBona explained that to him, it seemed Google was the place Computer Scientists go to practice Computer Science. Thus, though it would be nice to reach other people such as open source enthusiasts, their real desire was to reach computer scientist practitioners. The rate of innovative product development, focus on algorithms and performance, and more demonstrate that nurturing Computer Scientists is a key part of Google's success. Thus I found it very interesting to see the languages chosen by entrants in Google's AI challenge organized by the University of Waterloo. Read on the see the list.

The languages used for the contest include the following:

• GCC 4.3
• Java 1.6
• Perl 5.10
• Python 2.5
• Ruby 1.9
• GHC 6.8.2
• Mono 1.9.1
• Node.js v0.1.27-5-g1737cdc

Of these, the top 30 results used C++ almost exclusively. That's interesting! Despite the success of Java in industry, and rapid growth of languages like Ruby and Python as we highlighted in "What's going on with open source job opportunities?", it seems there's a preference for C++ for smart people to solve this AI challenge. Now, why might that be?

See the winners and more here. Congratulations to the winners and to the University of Waterloo and Google.