Hopefully this isn't too subjective. I am curious to know what measures to use to assess the quality of a given software release. The measures would need to be applied to multiple different projects within the organisation and so the measure have be reasonably generic.

Ideally I would like to use the measures to help move release deadlines when necessary as well as to push back to developers when I feel enough unit testing is not being done.

So far I have thought of the following that could be potentially measured:

• Defect rate per hour of test time.
• Defects per hour of dev time.
• % of defects that are coding errors.
• % of functionality with no outstanding defects (assuming the testing cycle is complete).