Estimates are broken. We've just fixed them for you

aboutfiveminutes.com

1 pointvrudyi2 comments

Never read and understood JP Lewis' "Large Limits to Software Estimation", eh?

http://scribblethink.org/Work/kcsest.pdf http://scribblethink.org/Work/Softestim/softestim.html

It's hard to outwit Goedel and Tarski on this stuff.

Amazing, thank you. I haven't read it, but definitely, will.

Obviously, there won't be a strictly mathematical solution. Neither optimal one. We can try to create something that is good enough for a limited set of problems, though. If it works at least in half cases, then I'm happy.

I don't think we can create a methodology or magic something. But having a repository of estimates for similar projects could be great.

Reply on news.ycombinator.com