Comments on: code reviews, tasks http://malvasiabianca.org/archives/2005/08/code-reviews-tasks/ Tue, 22 May 2012 01:20:09 +0000 hourly 1 http://wordpress.org/?v=3.3.2 By: David Carlton http://malvasiabianca.org/archives/2005/08/code-reviews-tasks/comment-page-1/#comment-3479 David Carlton Mon, 08 Aug 2005 16:53:31 +0000 /?p=132#comment-3479 We're doing some amount of measuring. Specifically, we're measuring how many hours we spent programming each week, how many stories we under- and over-estimated, how many stories we expected to finish but didn't. So we probably could establish an average underestimate percentage based on that. I'd rather leave that as a bit of a last resort: for one thing, I think we have more of an estimation problem on long stories than on small stories, which suggests to me that we have to be better about splitting stories. But if we don't get some traction on our estimates soon, I think that applying a blanket percentage makes sense. We’re doing some amount of measuring. Specifically, we’re measuring how many hours we spent programming each week, how many stories we under- and over-estimated, how many stories we expected to finish but didn’t. So we probably could establish an average underestimate percentage based on that.

I’d rather leave that as a bit of a last resort: for one thing, I think we have more of an estimation problem on long stories than on small stories, which suggests to me that we have to be better about splitting stories. But if we don’t get some traction on our estimates soon, I think that applying a blanket percentage makes sense.

]]> By: Ben Fulton http://malvasiabianca.org/archives/2005/08/code-reviews-tasks/comment-page-1/#comment-3477 Ben Fulton Mon, 08 Aug 2005 15:58:44 +0000 /?p=132#comment-3477 Are you going back and comparing the estimates with the actual amount of time? (Tracking your velocity?) You might be able to establish an average underestimate percentage and modify the estimates based on it. Are you going back and comparing the estimates with the actual amount of time? (Tracking your velocity?) You might be able to establish an average underestimate percentage and modify the estimates based on it.

]]>