Re: How to Estimate Project Time?

[Arthur Comings <atc -at- CORTE-MADERA -dot- GEOQUEST -dot- SLB -dot- COM>]

What a crock. Keep in mind that if you do give in and start flying
blind, you will get no credit at all if something goes wrong. It will
be your fault -- not the fault of the imperfect estimate you had to
make with insusfficient input.

The only time to do this right is with a trial run -- and the one trial
run you suggest is probably not enough. Now, *three* would really give
you some knowledge to work from. The only way to go when people insist
on an estimate when you don't have enough to go on is to say, "OK, you
forced me, but I have to give you something I can live up to. I've
never held a project up, and I'm not going to start now" -- and give an
estimate tht is two or three times longer than your gut minimum.

Estimating is an art, and some good things have been written on it. And
freelancers have got to get it down soon -- or go crazy. Since you're
in-house, you have time to learn.

Someone already suggested an hour per page. Maybe, since some of the
text is already there, but those procedures may hang you up. A day per
new page, start to camera-ready, is sometimes required. If it were me,
I would do enough of the trial run on my own to get a feel for it, and
then triple my estimate, which is always going to be too hopeful. If
they don't go for it, break the job down into all of the tasks that
will really go into it, and try to assign realistic times to them. When
you multiply them out, it gets impressive.

My experience is that, no matter how tough managers talk, what they
want is a realistic estimate that you can live with -- and deliver to.
And they'll swallow any dose of reality if you present it early enough,
before crunch time. Don't be cajoled into giving them something based
on hope. They'd like it, but they wouldn't like the results a few
months down the road. You'll be responsible for those results, and the
time it took will be forgotten.

Let us know how this comes out, and I hope you're on the Framers
mailing list. Yell if you need more help.

Good luck



Arthur Comings

GeoQuest
Corte Madera, California
atc -at- corte-madera -dot- geoquest -dot- slb -dot- com




> I suggested that we do one unit:  reference page, user overview, and 3 topic
> pages and judge time by that one unit.  The company feels that is not a
> reasonable approach.  They want an estimate without any trial run.

> Help.  Please do not be too tired from the STC conference to help me out.

> I will be working in Frame.  The basic paragraphs tags are prepared.

> The company will not tell me the maximum time they will allow before discon-
> tinuing the project.  How would you handle the estimating task?