|Perl: the Markov chain saw|
Re: How to calculate development time?by shotgunefx (Parson)
|on May 31, 2001 at 15:37 UTC||Need Help??|
If you are going to work on it only a few hours a day, I find that a huge time waster for me is getting back up to speed where I left off.
(Often I'll work a day or two straight on something complicated otherwise I waste half my time trying to remember where I was).
As far as calculating time (and cost), the job detail is pretty vague. Sort of like asking "How much is a car", there's a whole spectrum from the Yugo to the Bentley. I'd be very careful to state everything that you are and are not contractually obligated to perform and guarantee.
When I first started working for myself, I played it fast and loose. I've done a lot of different types of projects and my <nobr>"Guesstimates" ™</nobr> have always been really close... until.. I had a client that expected our 5 figure software to do everything that there aborted 7 figure software couldn't do. Do to the lack of a rigid spec, it was a painful ordeal. Luckily they had a burn rate on the order of a super nova so the problem resolved itself.
I know this doesn't really help come up with an estimate but I hope it gives you a bit to think about.
Whenever I am quoting a job, I mentally map it out into the base components and actions and think can I do this easily? Have I done it before? If the answer to the latter is no, I do a proof of concept real quick to determine the complexity.
When all else fails, come up with a reasonable and realistic guess of cost and time frame and then double it. :)
"To be civilized is to deny one's nature."