Dries' keynote

I guess I don't have to repeat that Drupalcon Boston 2008 attracted 800 drupalians. So eventhough the room for Dries keynote this morning was huge, it was packed.

Dries talked a bit about the new features in Drupal 6, we all know and love and mentioned that the download statistics prove that the Drupal user base is growing considerably. Inevitably, these new users are less geeky than the early adopters already on board and Drupal should accommodate for them, like for example a redesign of drupal.org.

Dries also lined out some plans about Drupal 7, based on
the survey he conducted in august 2007. The Drupal 7 killer release would consist of the 7 most requested end-user features and the 3 most requested developer features topped with an extra focus on usability. He showed a funny (or is it sad?) eye tracking video from the Drupal usability testing at the University of Minnesota Libraries of a user struggling with Drupal administration and gave the example of users taking 30 minutes for some common tasks that would take just 30 seconds for a veteran Drupal user.

Another issue he touched was the question "when will Drupal X be ready?". The survey learned that 65% of the respondents were happy with one major release per year. Drupal 6 had 5 months development and 7 months code freeze. Drupal 7 is already one month in development and not much time is left for implementing the ambitious goals. Dries suggested to change the development/freeze ratio to for example 9 months development and 3 months code freeze, but on the condition that Drupal core would get a 100% test coverage. The signal is clear: Drupal needs to embrace testing if it still wants to grow.

Dries mentioned however that the right testing frameworks still has to be figured out. Simpletest is already a popular option (e.g. check out the Lullabot articles by Robert Douglas and Angie Byron). However, Dries wants to ship the core test with core and Simpletest is a huge chunk of code, which is not a desired property for material to be included with core. So Drupal developers are a bit in the dark right now on how to start implementing tests the right way. I recently started to add some Simpletest tests to GeSHi filter for syntax highlighting, but I guess it's best to wait a bit until the uncertainty fades away.

In the last part of his keynote Dries presented his vision about how Drupal should fit in the future of the internet. Despite his dislike for the terms, he believes that after "Web 2.0" was about users and social networking stuff, "Web 3.0" will be all about data, like data portability and web services. The focus in Drupal should shift from the infinite extensibility we have today to the infinite interoperability needed tomorrow. That means for example that Drupal should be extended to output formats that are better machine readable than HTML like XML and JSON.

Another piece of the puzzle according to Dries is to give more power to fields and less to nodes, so that fields would become first class citizens. To expose these fields he believes that RDF would make new cool stuff possible. With RDF the whole web could become one big database. He showed an example of geolocation data fetched from Wikipedia combined (or should I say "joined"?) with event calender data from another source to create a Google Maps like view of the event calender. Pretty cool stuff done with just a couple of RDF operations (like queries in the RDF oriented query language SPARQL).
The semantic web promises all over again, but Dries believes we're now on an inflection point with the the right tools in place to make it finally work.

Once again Dries showed that his involvement with Drupal is more than just writing software. He has a clear vision on how the internet/the web will evolve and how Drupal should fit in that picture to survive or even empower those (r)evolutions.

Great, now I made my fancy buzzword-speedometer explode.