Nielson makes a fundamental error in treating usability as if it were a modernistic, mechanical problem with a clear solution that simply has to be discovered. A more useful approach is to recognize that what is evolving is a visual language, or, more accurately, a set of Web-centric visual and textual languages with a large number of dialects, constantly changing and feeding off one another. There will never be a single "correct" grammar for the Web; language warps to allow individual expression that "breaks" the rules while still being understandable. That's why Nielson comes off sometimes like my grade 6 English teacher, trying to impose a single, useless standard grammar.

I have found that usability (and design language generally) that works with current North American Web visitors might not work well in Europe, probably certainly won't work well in Japan, and might not work at all in Africa. What I expect to see is the continuing evolution of a variety of Web visual language dialects, as is the case with industrial and institutional design in various parts of the world.

Nielson has been lately describing a future in which all sites begin to look similar; he's half right, because the visual language continues to be created, with vocabulary tacitly agreed upon, but we'll also begin to see more clear differences based on culture.