How do you estimate development tasks? For most this is usually a process of guessing. Sometimes the guesses are accurate and many times they are not. One person I highly respect in the software development field is Joel Spolsky. He wrote an article explaining the process of doing evidence based estimates. This is very similar to one of the new features in a product his company makes called fogbugz (which I have not used). A bug tracking system does seem like an obvious place for this type of functionality to live. Especially since it should be one of the core tools used as part of your software development process. The idea of doing estimates this way is interesting though. It got me thinking, why don’t more estimates get calculated this way? I hope to be using a system similar to this in the future myself.