Unambigous parsing is certainly a step towards that, but I don’t think it’s enough — and it looks like the working group agree with me, as they’ve made 4.1.1 stronger than just ‘unambiguous parsing’ — albeit not quite as much as I would have liked.

Validity only has a defined meaning when the content is described in HTML. For example, what’s valid Flash, valid PDF or a valid Excel Spreadsheet, a valid Windows Media movie, a valid MP3 soundtrack? There isn’t anything relevant or realistic. But if you take the approach of non-ambigious, that is a concept that translates usefully into content other than HTML. Validity could mean something in largely text markup formats, but nothing in the realm of binary-encoded content. Yet unambigiously parsed means something in all of them.

The advantage of unambigiously parsed formats is that they are open to be easier repurposed - and that’s a key feature to making content accessible.

Perhaps many technologist types tend to assume that any given site will be re-developed and re-designed to make use of modern technologies; and that it’s therefore only necessary to develop for the technology of the moment, but it seems more reasonable to me to assume the opposite.