God damn it

[IHateMyJob2004]

So, I had to do an estimate with this other fella. We need to write a test simulator for some real world code (C++). There are 5 "programs" that compromise the collection of software. So he went and got the SLOC for each piece. We did an estimate for the simulator based on this information. Then we hand it to our boss. HIS NUMBERS WERE 14 TIMES HIGHER THAN THE ACTUAL "REAL WORLD CODE".

So, I now know why our estimate ended coming up with 7 people full time for 3 years which made no sense. I figured 3 people and a year .... maybe

I didn't question the guy. He's done this for 30+ fvcking years and its such a trivial task. WTF! I should have known better. If you want it done right, do it on your own. Now I feel like a god damned idiot. Well, at the meeting I am not going to point at him, but he damned well better take responsibility.

 

[EagleKeeper]

1) Ask him off line, how he came up with the estimate.

2) With experience, he may have some insight that could be passed on to you.

3) Always double check (sample) work if something seems off.

Simulation and test can be more expensive than the actual development.

 

[imported_Dhaval00]

Read Brooks' The Mythical Man Month... you'll get a pretty good idea as to why projects are over-estimated [maybe properly estimated]. A year ago, I used to think along the same lines as you do right now. A lot has changed and a multitude of projects have given me different insights into how people, systems, and time amalgamate. Set realistic expectations from the word GO. Of course, that doesn't mean you are wrong, but a guy with 30 years of experience surely must have *some* insight as to why he/she came up with the numbers. Is he/she a contractor/temp?

 

[presidentender]

Project I'm on was supposed to be three guys, nine months. The team lead graduated in May; we were supposed to have shipped in April, then May, and now roughly "whenever." The two of us who are left are rewriting most of his code, because it's all shiny and pretty and real-world useless. 30 years of experience might have made him good, and probably did, but it might have done nothing but reinforce antipatterns as acceptable. Ask him about it.