This is why
the semantic web is stillborn. Yes, you can apply meaning to objects, but what's to say the meaning is correct?
Google has reiterated that it doesn’t use the “keywords” meta tag in its search engine web ranking results. The company’s search wonk Matt Cutts was at pains to distance Google from any suggestion that the world’s largest ad broker did use the keywords meta tag in its web search system. “Google does sell a Google Search …
The meta tag can be used as a 'checker' to see if the contents match up with the analysis of the site content - much better than just believing what it said (which was the major cause of the abuse).
If the words are at odds with the site content, page gets marked down a few notches.
Meta keywords were like YouTube and Flickr tags 15 years beforehand.
People abuse them all the time too, but there's more of a community effort to validate them. If a video is advertised as being Britnay Spears but is actually pr0n, they soon get rumbled and the video taken down.
I can't see any practicable way for Google to do that with web pages. Very few people would actually complain and by the time you factor in Ajax, how can any bot really know what it's looking at?
The untrustworthy domains are...? (More to the point, who determines what domains are trustworthy?)
How do you prove that the results are actually from a trustworthy domain and not from, say, someone untrustworthy spoofing the domain?
How do you keep the inevitable authority or authorities needed for a trust network to work from becoming lazy, incompetent, and/or corrupt?
Having been condemned by every email filter known to man, I feel, for once, Google's pain. There was that one incident several years ago when a political twist, if you'll excuse the expression, made my name hot on the message boards. That I had obtained the name almost 50 years before made me a deity in smarmy political circles. Who knew 15 minutes of fame could be so long.
Keywords is a common but never official tag. The only official tag is title. The people who use the meta data sections of documents as well as the content sections, librarians, know that the official tag is subject (heading), but without a directed graph, subject itself is not quite right. However, a full text search never appealed to these folks anyway.
Rather than fuss about "keywords", the more interesting question, for Google and Microsoft is: Do you use, and how do you use "author", "editor", "printer", "creator" and "contributor" ?
And, while I'm at it, Google and Microsoft, the Personally Identifiable Information tags http://purl.org/pii/terms/ ? Sorry, shameless self-promotion, ignore that last question.
The untrustworthy domains are determined exactly the same way the advertising domains are determined by an ad blocker. If the list proves to be unreliable or inaccurate, subscribe to a different list.
This is not a perfect system by any means. Neither is the system employed currently by ad blockers. I don't know about you, but I'm not looking at any adverts right now, and the page content looks fine. This stuff isn't rocket science.