Web Mining, XML and SE integration of thoses ?

Nice interview with Mr. Laurie Lock Lee, principal knowledge management consultant with Computer Sciences Corporation in Australia. The last paragraphs are the better's one:

If you could command sites like Google to make one or two big changes, what might you require?

Clearly textual clustering or summarization, which is what we are seeing from companies like Vivisimo, Autonomy, Semio, and Verity. This technology on top of a Google-type search engine, which reaches even further into the rich data sources available over the internet, will make the web even more useful than it is today.

Are there major trends in WM standardization that you think people should be watching more closely?

XML is already the de-facto standard for improving the ability for machines to interpret textual repositories. I would anticipate that XML will become as pervasive as HTML. We in fact may still see a de facto standard emerge.

Hope so, because since 99, Xml seem to me is only use to produce sites from database, but not really to structure the web for a easier comprehension for machine interpretation of the pages produced.