When I'm looking at a new open source project one of the first questions I ask is "how good is this software?". Unless you know something about the reputation of the author or the code it can be remarkably difficult to tell whether something is good or crap. Caveat codor.

This is another reason why unit tests are so important. They're like a badge of quality. The fact that there are tests at all is a good sign. Run them, do they pass? Scan through the test code; does the coverage look good? The tests are usually a handy API example, too.

Mark and his feed parser are famous enough that I figured it was good quality code. But seeing 1882 tests pass sure is reassuring and it's quite interesting to read the tests.

For the bad side of this see the tests for pywebsvcs. Many of the tests don't work last I looked. It's not clear which tests are even useful. I don't mean to slag the pywebsvcs developers — they're doing great work — but I'd sure feel better about the code if it had a test suite I understood.

techgood
  2004-03-28 00:19 Z