On Mon, 2003-03-24 at 12:14, Chappell, Simon P wrote: > I have just completed an initial estimate for a Struts-based web application. I am > curious as to what estimation methods folks out there use for their initial > estimates. The revised estimate will be delivered after more analysis is performed. > > I am fed up with making wild guesses for initial estimates and wondered if there was > any way to take the small amount of information available up front (i.e. the number > of screens on the user's initial workflow requirements document) and extrapolate it > into an estimate. I understand that the margin of error is gonna be big and hairy on > this puppy, but it's an initial estimate created with a dearth of information, so > what do you expect? :-) > > What I did was: > > Take the estimated number of application screens: s=13 > Take the number of application user roles: r=3 > Figure which screens are used by which roles and total: sr=32 > Estimate the number of actions: a = sr * 2 = 64 > Application units of work: auow = a + sr = 96 > Total units of work (including admin screens): tuow = auow * 2 = 192 > Estimate an effort factor per unit: ef = 4hr > Programmer work: pw = tuow * ef = 768hrs > > Some assumptions included in these calculations include: > 1. An average of two actions per screen role. > 2. The administration facility of an application is about 50% of the work. > 3. The effort factor is a figure arrived at from knowledge of the programmer being > tasked and the number of new technologies within the project. > > Your thoughts please.
Did you factor in the time for writing/maintaining tests? > > Simon > > ----------------------------------------------------------------- > Simon P. Chappell [EMAIL PROTECTED] > Java Programming Specialist www.landsend.com > Lands' End, Inc. (608) 935-4526 > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] -- James Mitchell Software Developer/Struts Evangelist http://www.open-tools.org --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]