Bag of Spoons
Just off the A1(M)

Fri, 08 Apr 2005

Building a semantic website

You may have noticed from a few things about my site that I'm interested in the so-called Semantic Web. This is basically about embedding computer-readable information in web pages. Normal web pages using html are really just formatted text. Computers are not great at understanding text, which is why search results can be so wrong e.g. 'bass' can be a fish or a musical instrument or a fat singer or that thing that makes your windows rattle when you turn the stereo up.

I've added a few things to this site that give it more 'meaning':

Location - my geographical location is embedded in the site header as a latitude/longitude. This allows sites such as GeoURL to show where I am relative to other sites with that information. Maybe, one day, search engines will use this to allow searches for local services. Just using names, e.g. Bedford, is not so reliable as there can be multiple places with that name.

FOAF - the FOAF Project proposes a standard file format to give others information about yourself. The actual file is in a form of XML called RDF and is not that readable, but it can be extracted to nicer formats as you can see by clicking on my name at the top of the page. My file has information about who I know, what I do, where I've been and more. I intend to add more information. The 'who I know' bit has more potential if those people also have FOAF files so that our social network can be extracted.

XFN - this is a way to add meaning to links. By adding some extra parameters to the html you can indicate if the owner of a page is a friend, colleague, relative etc. This has similar implications to FOAF for building social networks.

There are more potential options including those that add more categorisation to articles. See my semantic links. There's a few reasons why this has not really taken off. Ignorance (many have never heard of it), apathy (why should I need that?) etc. Some have written about why they think it is not going to work. There are the usual internet risks of fraud to get visitors to sites. But there are many sites using semantic data. Simon has built a wiki that includes the locations.

One of these days I will get around to writing some applications to make use of this data. Python has tools for the job. I have some ideas for practical applications. Instead of using Multiply to link to friends I could use the information in my FOAF file. Users in there could request a password to allow them access to certain parts of the site. This could include thing like my DVD library that is hosted at DVD Lender. The idea is that I could take back control of my own data instead of relying on other sites to look after it and having the risk of losing it if they disappear. The other thing I want to do is to generate the details that appear on the right hand side of this page from my FOAF, so the FOAF would act as my information store. Other people seem to be working towards the opposite effect and generating FOAF from web pages, but I'm not so sure about that. I still have a lot to learn about this and the subject is still in flux generally.

[16:31] | [] | Comments | G
blog comments powered by Disqus



< April 2005 >
      1 2
3 4 5 6 7 8 9

Subscribe if you like
XFN Friendly

Get Firefox!
Flying Spaghetti Monster

Hosted at VeloceSystems