Wednesday, June 24, 2009

Stuff I Have Found Interesting Today

Be The Change You Wish To See

Repeating good advice never gets old.

Requests to the Right Ear Are More Successful Than to the Left

Not very scientific, but it sounds pausible.

50% Time

I think to be great you must want it bad enough to dedicate this much free time to it.

The Funny Thing About Waste

A fairly long article about waste and how it is all perception. It ends with one of my favorite quotes "The greatest trick the devil ever pulled was convincing the world he didn't exist." that I can now atttribute to the author, the French poet Baudelaire, as opposed to the movie, The Usual Suspects, I remember it was used in.

QTB: Agile Adoption - How to stuff it up

Good notes on a talk by Martin Fowler on Agile adoption.

Control Charts

A new thread that starts with a good question, but then lays down the gauntlet with " may be time to drive a stake through the heart of Control Charts and Common Versus Special Cause of Variance in software." which brings thoughtful response 4173. This response has three links:

  • Control / Capability Charts on a Kanban Software Development Project - shows how he uses control charts in his team: I don't buy it, but if it works use it
  • An Introduction to Systems Thinking - which states "It is central to a systems approach that capability measures are used to measure performance...A key aspect of capability measurement is statistical process control (SPC)." This is straight Deming, but I don't think Deming would use them for software as you wind up measuring apples versus oranges.
  • Freedom from command and control - basically says targets are a bad idea, but capability metrics will set you free. I think measurement is a tool that is useful to know, but using it as a management tool is wrought with so many perils as to be not useful

I think control charts are a useful tool when faced with a process in which you can measure similar widgets. I don't see them as a useful tool in measuring or helping the overall software process as one chunk-o-work isn't really like another chunk-o-work. I have heard attempts to regulate the sizes, but then you are twisting the process to fit some preconceived goal (e.g. we must use SPC). You need a process, but it should be based on your goals, not your tools.

No comments: