Python at Google

A writeup of Greg Stein’s keynote at the SDForum Python Meeting gives some insight into Google engineering process in general and the usage of Python within the company in particular. Here is a summary of the summary.

• Three “official” Google programming languages: Python, C++ and Java
• Google likes SWIG and RPC
• Python used for:
  • Google build system: Makefiles etc.
  • packaging: internal packaging format like RPM
  • binary data pusher: distribution of data between lots of servers
  • production servers: monitoring, restarting, data collection
  • reporting: log file analysis
  • misc: some services such as Google Groups and code.google.com

Another interesting point is the approach to tackle problems and design new applications. I’m not sure if the quote is directly from Greg Stein, but it is consistent with other statements of Google (or should I rather say: unintentional information leaks in PowerPoint presentations).

When programming in Python, one should design for scalability. If you design properly, then you should be able to just throw more machines at the problem and your app should scale fine.

Google’s philosophy to new product development is to initially ignore any constraints such as CPU power, storage, bandwidth, and monetization.