tag:blogger.com,1999:blog-36118108.post4510590811426051165..comments2023-12-06T00:17:28.519-08:00Comments on Creative Chaos: Test Estimation - VMatthewhttp://www.blogger.com/profile/05956714498778698672noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-36118108.post-74546407699902055352010-09-15T05:30:19.293-07:002010-09-15T05:30:19.293-07:001) hwtsam.com - keep reading. You're getting ...1) hwtsam.com - keep reading. You're getting ahead of things. :-)<br /><br />2) Chris - yes, that's pretty much what I'm advising here, and I've had close to the same results that you did with that method.Matthewhttps://www.blogger.com/profile/05956714498778698672noreply@blogger.comtag:blogger.com,1999:blog-36118108.post-74717069425563746162010-09-15T00:51:12.789-07:002010-09-15T00:51:12.789-07:00I always found that managers were unwilling to wai...I always found that managers were unwilling to wait the amount of time it would take for a thorough estimate of testing effort. My back-of-the-envelope method was: find out how much time the development director estimated for coding, then take half of that and submit that as my estimate. Sometimes my manager would not be pleased with the number that produced, and ask me to cut it down.<br /><br />However, once the project was finished (over time and over budget), I recalculated using the actual results, and found that, as near as dam*it, my testing effort was within 10 days of 1/2 the development time as it had finally ended up.<br /><br />I don't believe this is infallible. No estimation method is infallible. However, it has several advantages:<br /><br />First, it's easily explainable, even to Dilbert's pointy-haired boss. One sentence will do it. <br /><br />Second, it's easily adjustable during the actual testing effort. If the programmers say that they will be four weeks late, you then add two weeks to your effort. When the pointy-haired boss comes back and says, "We need it earlier." you can then negotiate which features will be cut out in order to make that deadline.<br /><br />Third, once the manager accepts that this is your rule-of-thumb, s/he can do the sums himself/herself and cut the production cloth to fit the testing.<br /><br />This probably won't work for everyone, I'm afraid. It works for me so that's why I use it.Chrishttps://www.blogger.com/profile/06534842755063770798noreply@blogger.comtag:blogger.com,1999:blog-36118108.post-29864121524491760442010-09-14T13:22:22.115-07:002010-09-14T13:22:22.115-07:00This is hard to follow for me, because I don't...This is hard to follow for me, because I don't belive in "the cost of testing" - I believe in the cost of quality. There are so many factors that go into the role (and effort) of testing in that equation (even on the same team working on a similar project), that I think estimations are pointless.<br /><br />Now, I realize that I, for better or for worse, live in a world where we have smart managers who know this - but rather than figure out how to work with dumb managers, shouldn't we focus on making them smarter so we don't have to explain stuff like this?Anonymousnoreply@blogger.com