David Emery Online

Thoughts on BarCamp

20 February 2007

Due to the lack of a decent internet connection at Future of Web Apps my round up of BarCampLondon2 is going to end up being a bit briefer than I originally anticipated; I was going to try and write it during the day but it just never quite happened…

One of the first big themes of the weekend was of RDF Vs Microformats, and the semantic web – my conclusion is, quite frankly, RDF sounds interesting from a technical point of view but is never, ever going to take off. There is just no compelling reason for using it, other than because it’s interesting – that’s just not enough for most people.

Even Microformats, which most people at the conference seem to have taken for granted as a “success”, still don’t have a huge uptake – for exactly the same reason: as yet, there’s no really compelling reason for the end user to care about Microformats.

For crying out loud, most people still don’t get RSS…

The two most common demos for Microformats are converting an event listing into an ics file, and converting some contact information into a vcard. Both are nice, but generating vcards and ics files isn’t hard in the first place, and if your content is in a CMS it’s just a matter of a different template…

When we get built in browser support – which we’re going to get in Firefox 3.0 – it could get a lot more interesting, but until then Microformats remain a interesting curiosity.

The other big theme was OpenID, and I’m afraid I have similar reservations about that too. OpenID provides a way to have a single sign-in for multiple sites, which you control. In theory, it sounds great, and in fact I’m really annoyed that it isn’t as I’d really like to use it.

However, OpenID suffers from a pretty major phishing flaw which in my eyes means it’s almost completely useless. Basically, if you ever try to login to a “bad” site (or a compromised “good” site), they can very easily put up a fake page that gets you to enter in your OpenID username/password, which they can then use to login to all of you accounts. The only solution to this seems to be to force people to either stay logged in all the time, which most people don’t seem to like doing, or to get them to manually log themselves in by visiting their OpenID provider by typing in its address – completely disrupting the site login process.

Because of this, I can only see OpenID being useful for throwaway accounts such as blog commenting, where it doesn’t really matter if your account is compromised, which is a real shame.

Other than those two themes, Tom Coats did a very interesting talk on Social Networks which I’m still trying to digest properly, and Simon Willison – who was also the main OpenID cheerleader – chaired a very interesting panel on Django, which I think I’m going to investigate further.

My discussion on the future of music went really well I think – lots of interesting thoughts on music distribution and discovery to think about. Hopefully, tomorrow I’ll most some of my thoughts on Future of Web Apps; it’s been very interesting to compare and contrast the difference between it and BarCamp London, more on that then.