I have following baselines available, as examples
1. Defect distribution data for development projects say as 20% are found in Analysis docs, 20% in Design docs, 40% in Coding and 20% in Testing.
2. Document defect density = 0.5 / page
3. Code defect density = 10 / KLOC
Now is following approach reliable and correct
I first estimate for LOC and then using code defect density estimate total coding defects.
e.g. 100 KLOC estimated, so predicted coding defects = 100 x 10 = 1000
Now this will be 40% of total defects which means I can expect total (1000/.4) = 2500 defects throughout entire project. From this and using % distribution, estimate analysis doc and design doc defects count. Now using document defect density predict the size of these documents. i.e. use % Defect distribution to in turn arrive at document size.
Or is it more practical to understand requirements in detail, identify sections of documents, TOC and then estimate for document size independently.
Life should NOT be a trip to the grave with the intention of arriving safely in an cool and well preserved body, but rather to skid in, chocolate in one hand, beer in the other, body wrecked, totally worn out and screaming WOO HOO what a ride!
I have never heard of or come across estimating the size of a document. In 30+ years in IT I have never had the need to estimate the size of a document.
Estimates that I am familiar with are to estimate the effort it will take to complete a task. If you are attempting to estimate the effort it will take to accomplish an analysis or design document I would not be using LOC.
Code is developed based on the contents of these documents and until these are complete a reasonably accurate estimate on LOC cannot be done.
The size of the document is what it is and the main concern is that the documents contain sufficient information for the developers to code the application and for the testers to test the application.
Estimating effort is worthwhile as you need to understand the effort it will take to develop an application - estimating the size of a document is a useless waste of time and effort.