« New blog: Content Thing | Main | Victorian accessibility guide »

Death of keywords

Danny Sullivan from Search Engine Watch reports on the death of meta tags. To quote:

Now supported by only one major crawler-based search engine -- Inktomi -- the value of adding meta keywords tags to pages seems little worth the time. In my opinion, the meta keywords tag is dead, dead, dead. And like Andrew, good riddance, I say!

To me, this really highlights the challenges (futility?) of the so-called "semantic web", where everything describes itself, cross-linking happens automatically and accurately, and search engines only return useful results...

If we can't get even simple keywords tags to work in practice, what hope is there for RDF, and the rest?

[Thanks to IDblog.]

Posted by jamesr on October 11, 2002 04:11 PM
Categories: Information architecture, Search tools, Web development

Comments

The problem with simple keyword tags is just that they are to simple and do not provide an acurate description of the content required by those searching.

Posted by: amateur on October 16, 2002 09:08 AM

To me, the point is that you can't trust people to be honest about their site descriptions. In many cases, it's simply not in their best interest...

It sounds harsh, but that has obviously been the experience of the search engines.

Posted by: James Robertson on October 16, 2002 09:56 AM

I think amateur hit the nail on the head - using keywords isn't really much better than using page-scraped keywords, in fact the latter may even be better for the reason James refers to.

But I really don't think the failure of keywords has much relevance to the semantic web. On this page there is a syndicate button, and I've just added the link to my Amphetadesk newsreader. The posts on this page have metadata - title, description, date etc so my reader will present them to me in an appropriate fashion.

The metadata is being created and read with very little user input - I had to look at the source to find that it was RSS 0.91. This uses a few commonly defined terms to convey the semantics, but RDF (such as RSS 1.0) allows virtually any meaning to be shared.

The kind of information that could be used in tools like search engines can be conveyed in exactly the same fashion. If there had been a category element in the RSS feed, then that also could have been used in the reader (assuming it understood the element). The amount of human work involved in putting other classification metadata in needn't be any more than that involved in putting the date in - i.e. zero. A smart authoring tool would let the writer place the data in their own information system in a retrievable fashion, and relevant parts of that metadata could be transferred to other readers.

Posted by: Danny Ayers on October 26, 2002 04:07 AM