I'm looking for open freely available data sets related to software development quality: code quality, defect rates, comparison across programming languages. Where/how can I find such data sets?
Here are a few sources to take a look through:
Most comprehensive and closest to what I think you're looking for is a project by the acronym of PROMISE (PRedictOr Models In Software Engineering). Go to http://promisedata.googlecode.com/svn/trunk/defect/ to go straight to their data which is organized by code base and into CSVs. For example, tomcat.csv. It looks like the project also has a number of tools to help analyze and create models from the data.
Eclipse Bug Data! is, as it sounds, a project to mine bug and version datas from the Eclipse project.
NASA also publishes "cleaned versions of publicly available NASA MDP software defects data sets" over at http://nasa-softwaredefectdatasets.wikispaces.com/
EDIT: there are also a number of academic papers on this topic that you might want to skim to see if they reference specific data sets or sources.
Note: This answer only covers a particular aspect of code quality.
OpenHub (ex Ohloh) contains 666,192 projects and mentions for each one:
- The number of comment lines
- The total number of lines
- The programming language used (or proportions if several used)
So, you can use that data to compute the comment/code ratio for many programming languages.
Some of the projects listed under the data sets section of dspinelli's awesome list for mining software repositories (MSR) stuff might be useful. Some examples (additional information found on the GitHub list and through the links themselves):