phranque - 1:08 pm on Dec 31, 2008 (gmt 0)
by semantics, i mean most of the discussion in the threads linked from this WebGen post:
A tool like this?
Semantic Data Extractor
i like that tool and use it often but it only tells you what is semantically "there".
it also be interesting to measure the semantic "noise".
even a metric such as percentage of semantically defined textual content could be useful.
something that graphically shows semantically defined content in contrast with other content.
maybe even something that could determine whether headers and tables are used for proper semantic purposes.