Predicting Project Time

Hello,

Short Version
What tips, tricks and techniques are there for making more accurate predictions of the time it takes to complete complex and challenging programming projects?

Longer Version
I’ve been programming now for 7 years and have a degree in Computer Science BSc, but I still find it quite hard to accurately predict the length of a programming project! Especially if it is something challenging or using a new programming language/technology. One of my friends is a Project Manager in a big IT company and says that developers seem to get better at predicting projects as they get more experienced. I have found that true to an extent so far, but I can still be quite wildly out in my prediction of how long a project will take! She also says that although a deadline is important, generally you have to choose between flexible objectives or flexible deadlines.

I suspect that challenging projects are inherently complex and unpredictable, you often see big IT projects over-running. BUT I wonder if the community has any tricks and tips on improving prediction accuracy.

Here is an old technique I used once which claimed to have success:

http://en.m.wikipedia.org/wiki/Function_point

It is a lot of work though.

Sent from my XT316 using Tapatalk 2