Tag Team
An intriguing Flash-based thought experiment, created by Matt Biddulph, sets out to illustrate how del.icio.us could be used to enrich existing websites. The demo is undeniably cool stuff, but my first reaction was probably not what was intended. The BBC clearly has the resources to do consistent, ubiquitous semantic tagging in-house. Wouldn’t it be even cooler to be able to exploit these professionally-authored tags alongside those created by del.icio.us users?
It would make a lot of sense for the BBC and similar organizations to annotate their pages with tags manually and provide search functionality to browse their websites by keying off the relevant tags. This would provide the same functionality as that shown in the Biddulf demo, but with much greater precision. All that would be needed is a way to embed tags directly in a webpage rather than storing them on a separate site. And what do you know, this is exactly what is provided by the HTML meta tag. If you view the HTML source of the BBC homepage, for example, in the document header you’ll see the line:
<meta name=”keywords” content=”BBC, bbc.co.uk, Search, British Broadcasting Corporation, BBCi” />
Using this approach, tags could be placed in each content page and viewed using the del.icio.us interface, which could be embedded in the page as in the Biddulf demo. Voila! The semantic web is now a reality.
Why this hasn’t happened already is explained well in this Many-to-Many post by Clay Shirky. I think Clay is spot-on in his assertion that tagging must require little specialized knowledge if people are to make the requisite effort. But this still doesn’t explain why large organizations like the BBC don’t tag up their pages, since they could easily develop an internal tagset for their specific domain that could be used with minimal effort. This is more likely due to the lack of convenient ways to query these tags using existing search engines. Google’s Advanced Search, for example, offers options to search anywhere in a page, in the title of a page, in the text of a page, etc., but not in a page’s keywords. The result is basically a Catch 22 situation: people don’t tag pages because there’s so little benefit from doing so, and search engines don’t bother to offer querying via tags because so few pages are tagged usefully.
Del.icio.us is perfectly positioned to bootstrap this process by including the tags present in a webpage’s HTML meta element in its bookmark’s tags so I can always access the bookmark via the tags that the page’s author chose, if any, along with those chosen by the community at large. Already this would provide considerable motivation for folksonomy-friendly folks to tag up their pages. This in turn might induce search engines like Google to offer a del.icio.us-like interface for the entire web, based on these tags. Del.icio.us could then use Google’s web service API to fold this content into its lists of bookmarks, so you would see the top Google matches for each tag alongside standard del.icio.us content.
This is the best illustration that has occurred to me so far of why there doesn’t have to be the tension between folksonomy and more formal tagging that many seem to be taking for granted.
5 Comments »
Trackback URL RSS feed for comments on this post. TrackBack URI
Leave a comment
Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>
I don’t think the BBC has the time or resources to add tagging, especially since it’s still at the “internet fad” stage. They’ve got to justify everything they spend these days, and money is tight.
Anyway, this simplistic tagging scheme has serious bugs. It doesn’t deal well with synonyms, f’rinstance apple and windows.
Comment by scruss — 2/3/2005 @ 5:15 pm
i hope this happened if it is that good thing. but i gether that searchengine like google are not weighted meta tag at all because seo abuse. so, many people are not bother to think meta tag keywords. i like a combination of liquid (by http://www.liquidinformation.org/index-fr.html) +delicious (demonstlated by http://www.hackdiary.com/misc/firefox-delicious-demo.html) +Wikiproxy (demonstlated by http://www.whitelabel.org/archives/002248.html) wrapper, so that one word are possess more than one tag to refere or link.
Comment by huxing — 2/3/2005 @ 5:22 pm
Thanks, huxing. Interestingly enough, I just read an article by Google’s director of search quality where he confirms exactly that: they generally don’t use webpage metadata because of the potential for gaming. However, I don’t think this is a problem for the kind of scheme that I am proposing. What I would do is to take advantage of the standard meta element scheme attribute to indicate that the keywords are specifically designed to be used for del.icio.us-style tagging, like so:
We don’t really care whether Google takes this information into account when doing its page ranking, but we do want to be able to query on this data if we ask to do so explicitly. In other words, some unscrupulous person could put a bunch of inaccurate tags into their page, but they wouldn’t have much to gain since the order in which pages are displayed when someone searches for these tags would depend on the normal Google page rank, which would be unaffected by this choice of tags.
Comment by Matt — 2/3/2005 @ 6:37 pm
You are posting comments too quickly. Slow down
Comment by cep program — 5/14/2008 @ 8:50 pm
thank you nice sharing
Comment by cep program — 5/14/2008 @ 8:53 pm