Re: What is quality?

[Typo? What tpyo? 04-Jan-1994 1238 <jong -at- TNPUBS -dot- ENET -dot- DEC -dot- COM>]

Jonathan Leer commented on documentation quality and asked if anyone has had
much success tracking document performance.  I would like to draw a distinction
(which I thank Saul Carliner for pointing out) between evaluating and
predicting the performance of a document.  Evaluation is after the fact (or
producing and distributing the document), while prediction is before or during
the fact.

Examples of document evaluation are reader comment cards and usability testing.
Examples of predictors are documentation metrics such as the number
of graphics, headings, and index hits per page.  These metrics ought to
correspond to your customers' critical success factors; in other words, you
should measure things you know your customers want to see.  (Your mileage may
vary.)

Both prediction and evaluation have a place and have value. Evaluations are
inarguable -- when they are statistically valid, but they're never
statistically valid, and usability testing is quite expensive and difficult.
Doc metrics are indirect and seemingly crude measurements (and collecting
customer CSFs is subject to the same problems of statistical validity), but
they are fast, cheap, and easy.

I have done quite a bit of research on document quality and metrics,
if anyone is interested in continuing this thread.

                -- Steve