Kris Khaira

I’m right now at the last keynote of DrupalCon Paris 2009 – Semantic Web Fundamentals by Dan Brickley who is quite active in the W3C community.

Who is Dan Brickley?

• Accidental standards-nerd
• Making things in the Web since 1994
• Found metadata community & W3C through trying to make better Web sites

Notes:

• Web 2.0 mashups related to semantic web – Google Earth / Earthquake maps
• Layers of data – What links the layers?
• Linking – Working on ways to identify when two documents mention the same thing and how to tie them tgether.
• A bit of history – Dan’s showing the first document of the web by Tim Berners-Lee
• Different layers of how info is used should be linked sensibly if info is common e.g. different websites talking about the same school
• Gmail – things are linked together e.g. threading
• W3C POWDR – label pages or collections?
• All LCSH topics on Paris – why can’t they be associated?
• Shared datasets in networks of websites e.g. universities with libraries, museums, government departments
• Yahoo SearchMonkey – a way of of customising search results
• Adobe XML – stuffing RDF descriptions inside files. Been doing this since 2001 with TIFF JPEG, PNG, GIF, PDF
• Drupal Fields API – attach arbitrary properties to anything in Drupal 7. RDF does the same for the Web at large. It’s natural to bridge the two.

Drupal & Data APIs – SPARQL is to RDF as SQL is to RDBMs. SPARQL scripting looks very familiar to anyone working with SQL.

There’s a growing ecosystem around linked RDF data including people. Amnesty International’s dev team is getting involved.

Q&A

There’s a potential for manipulation when people put their info into RDF. What kind of manipulations are already appearing?

We have to expect lies because the Web is full of it, like with HTML. Build as many links to real data as possible. Based on those links, we can filter away fake data.

How do we manage links between the computer and the real world; and the data on our website?

It’s natural for people to make different decisions at this point based on their needs. For example, my info is scattered on different websites e.g. blogs, Flickr. If you have a clear linking model, it doesn’t matter much.

Is there an initiative to collect taxonomies in specific industries e.g. leisure and engineering. Basically a way to register your taxonomy so others can use it?

Good question. We don’t have a nice directory but with things like SearchMonkey, people can choose to see what’s out there to use. That’s something that we’re working on at W3C – some way that’s easy to maintain.

With regards to privacy, is there a way to index my text but not let everyone to view them?

RDF is agnostic and doesn’t care where your data lives. What we found is by putting public data on a public network, we get a huge network of useful data. [NOTE: you can always probably use Drupal’s access control to control access to certain pages for a group of users.]