Matts Ramblings

So, I’ve been meaning to RTFM on agile for a while, and today I ordered a few books from amazon:

• “Agile and Iterative Development: A Manager’s Guide” and
• “Succeeding with Agile: Software Development Using Scrum”

Hopefully they will arrive soon. If Anyone has any other book suggestions, please let me know!

In the meantime, I’m making notes of the things that I’m not expecting Agile as a process to solve. Some of these things are quite important, if not critical. To start with, theres still two obvious questions which so far I’ve not worked out how to answer:

• When will you be finished, and
• How much will it cost

Im not really expecting a process methodology to answer those questions, at this point I’m leaning towards a probabilistic method of answering both questions. Which could be useful down the track for showing how changes impact both the overall cost of the project and the overall end date.

Actually, let me extrapolate on my thoughts on that. Lets, at the start of any project, assume that any long term estimates are completely incorrect, and that over the duration of the project estimates become more accurate the more well defined the project, and the tasks that have been estimated, become. They will however, never  be completely accurate.

Adding new functionality in the middle of a project, whilst completely allowable in an agile process methodology, is likely to introduce error back into our estimates as downstream dependent tasks are now less well defined.

Perhaps there is a way to track the probabilistic error increase?  Perhaps we can graph the uncertainty, and track it against overall progress, bugs, errors, and most importantly cost. The answer to the question of “when will you be done” becomes another question “when will you stop changing things?”. Only this time, we might have a graph to back ourselves up.

Ive got no maths, stats or otherwise to back this up, its just a hunch for now, but its an interesting thought nonetheless

A simple tome, but one who’s importance should not be overlooked. The most important phrase in software development is, simply:

what problem are you trying to solve?