Quality software

Here’s an article on Ars Technica about the differences between Microsoft and Google in code testing:

It’s a real world example of the usefulness of writing tests that Zed’s books are always pushing us to do.


It’s definitely about cadence. Big Bang releases are almost always that. They are great for marketing purposes but do not de-risk software delivery or improve quality.

A frequent delivery model is much more sustainable, quality improves as the chunks are smaller and when errors do sneak through to production, you have the mechanics in place to address it very quickly.

This is the new view of quality. It’s not about error free, but about fixing the errors that your customers perceive as the most impactual.

Having said that, there is no excuses for developers writing code that logically and functionally does not operate at unit level.