BookReview: Text, ConText, and HyperText
by Edward Barret, ed., The MIT Press, 1989

A collections of essays about writing with the computer and using computers.

[p176] Essay by Edmond H. Weiss, Usability: Stereotypes and Traps [...] First, testing a manual only after it is a complete draft is much too late in the life of the publication. By the time a book has reached the draft, its most serious failings are deeply ingrained--``wired in.'' It is much too late, and probably much too expensive, to correct them.

Second, testing the pieces or ``modules'' of a larger publication does little to assuer the usability of the publications as a whole. In just the way that succesful ``unit tests'' of a computer program do not betoken an installed, functional system, neither do succesful tests of the parts of a manual prove that the book will be usable.

If we pursue usability as thought it were an abstract, technical imperative, then we confuse usability with perfection--the same mistake many quality enthusiasts make. Because there cannot be a ``perfectly usable manual,'' we soon see that what is really meant by a usable manual is one usable eenought to meet our business or management objectives. And a while later, when costs become an issue, usable enough to please readers and customers we most wish to please.

"The On-line Environment and In-House Training" by Barrett and Paradis, is a good article.

Via Rob 1990