Re: estimating page count for SW user guide

["Anthony Markatos" <tonymar -at- hotmail -dot- com>]

Debbie Pesach asks:

How do you estimate the page count for a new printed software User Guide 
(that isn't based on any existing documentation)? I'm looking for some ways 
to estimate my project.

Tony Markatos responds:

I create data-flow diagrams to "chunk" the system into tasks.
The data-flow diagrams tell me if I have properly partitioned the system: 
each task should not have more than four or five inputs and outputs (total 
combination).  (Note: These inputs and outputs are also called interfaces.)

I know, from experience, that, if a task has four or five interfaces, I will 
need to create about a page or two of procedural text (not including any 
systems overview material). If less, usually a page or less.

My estimate is now a matter of simple arithmetic.  You will be surprised at 
how accurate such an estimate is (and therefore, how comfortable you feel 
with it).

Debbie Pesach says:

[The author of a well know book on book managing documentation projects] 
..recommends breaking out the various tasks that will be documented......

Tony Markatos responds:

Lets stop here for a second.  How are you supposed to "break out" the tasks? 
 Only data-flow diagrams rigorously GUIDE you through this all-important 
activity.  Such guidance is very important for new products and larger scale 
projects: In such situations, just identifying what the tasks are is a very 
daunting task - it so easy to miss important stuff.

Debbie Pesach continues:

...and then deciding on their relative level of difficulty.

Tony Markatos responds:

How do you decide a task's "relative level of difficulty"?  By "gut feel"?  
For new products and larger scale projects, prior experience can be of very 
little value.  And "gut feel" has severe limitations.

Using my (above mentioned) estimation methodology, the relative level of 
task difficulty is very formally determined - by evaluating interface 
complexity.  You will find that such a rigorous formal analysis can make a 
tremendous difference in terms of the accuracy of your estimate.

Debbie Pesach says:

Then [using the popular methodology] you are supposed to create a thumbnail 
sketch of a sample chapter section to determine an average page count needed 
for an "average" task. Multiply this by the average number of tasks, add in 
other content (TOC, index, front matter, etc.) and you have a general book 
estimate.

Tony Markatos responds:

Again, this is all based on some sort of touchy-feely guess.  No rigorous 
formal calculation comes into play.  And for new products or larger-scale 
projects, such an approach soon becomes very shaky.

Tony Markatos
(tonymar -at- hotmail -dot- com)

______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com