We all know it too well - meanwhile for a long time!
Still today (2014-MAY-09) the latest specs-draft specifically warns:

"the outline algorithm cannot be relied upon"
(source: http://www.w3.org/html/wg/drafts/html/master/sections.html#outlines)

To me it appears as if there are lots of exceptionally skilled professional craftsmen trying to cover the roof of a house at a time when beams carrying the roof aren't put in place yet.

What the heck is going on?

Please excuse my total incompetence in the specific field of establishing a new spec, but - and I believe many agree - that this process cannot go on forever without running the risk that nobody will ever apply it once (if) it ever gets done.

Please establish and commit a solid outline (foundation of the house) so that we can start to build upon. At this point, any debate over the flower colors decorating the house is futil.

Cement the semantics (!) by taking a few grand news papers as a base (with dated, static text and images), then watch a Harry Potter movie and observe a book reading scene where images become alive (video = images and sound)) and sometimes them living images become the continuation of the story or even lead to new stories. I'd assume that the described spectrum covers some 85+% of the use cases.

Today we can only be sure that the outline of a book by its table of contents will probably remain in the spec, anything other is subject to change (=varying the foundation) leading this consortium to pronounce that "the outline algorithm cannot be relied upon"!

So what are the very good reason(s) for taking the effort to apply proper semantics today?

Please be honest the answer is 'practically none' and 'theoretically many' for the simple reason stated by this consortium: "the outline algorithm cannot be relied upon"!
(q.e.d.)