Comments on: agile, top down

By: kevin

kevin — Mon, 01 Aug 2005 12:06:20 +0000

Ah! So it’s the tested of “running tested features” that is the key. If it ain’t tested, it doesn’t ship. Or if it does, it doesn’t count to throughput. Cool.

And by further implication, I guess that any untested part of a feature must disqualify that whole feature. Nice.

By: Jason Yip

Jason Yip — Sun, 31 Jul 2005 23:10:32 +0000

The important thing to note is that “crap” does not contribute to throughput. The Goal had a reference to this when the main character asked about dropping testing.

By: kevin

kevin — Sun, 31 Jul 2005 21:13:34 +0000

I guess you’re right, 3-4 months is a long time. I was thinking of the kind of shop that is already working in iterations, but just not doing testing and stuff. So they’re churning out crap at a really fast rate. Adopting TDD, pairing and the others will slow them down, possibly until the new practices start affecting the incoming defect rate – ie. until customers have started using the new stuff in anger. Only then will they recover the time invested in quality practices, it seems to me…

By: Jason Yip

Jason Yip — Sun, 31 Jul 2005 01:16:10 +0000

Ron’s article was specifically assuming that you were doing this as an executive or highly-placed manager. I would agree that coming from a different place, the advice would be different.

I’m also skeptical about the 3-4 months drop of throughput. Ron’s saying measure RTS and then focus on improving that. Just doing the measurement wouldn’t affect RTS. All that does is tell you what it is. If we’re really saying the other stuff isn’t a valid measure of throughput and they drop…

The changes to improve the RTS might cause a temporary dip but 3-4 months seems a bit long.