On Mozilla and Thunderbird

July 27th, 2007

I wanted to point out that Brendan Eich from the Mozilla Corporation left a thoughtful comment on my post about Mozilla and Thunderbird.

He mentions about Mozilla, as platform and the work with Joost, which also incorporates the use of RDF. I was critical last year when Mozilla dropped use of RDF, though I've since become more sanguine about the issue.

The whole comment is a good read, but Brendan did make a comment on the perceived 'pragmatism' of the decision to find new leadership for Thunderbird I wanted to repeat here:

Pragmatism of the Mozilla kind rubs many people the wrong way. Why shouldn't we invest liberally in various dark horses and lost causes, such as our old RDF code or an upgrade to it, and even take the footprint (download and runtime) hit of including such features? I think the answer has to be stated bluntly: because we'll fail if we do that, and then the Firefox lever against Microsoft and anyone else with large market power and little concern for the Open Web will have an easier time.

Comments
1
Danny - 2:55 am 7/29/2007

Hmm, I'm pretty sanguine now too, but if one of the reasons for Moz dropping RDF is to be able to compete with MS it's a little short-sighted. Ok, sure, SQLite integration will enable a lot of new & interesting apps/extensions to easily be built on the platform. But the relational model is seriously suboptimal for representing web-shaped data.

In contrast, Microsoft have been moving towards an entity-relationship model in their Entity Framework, which is a much better fit, especially when you look at their their Astoria project, which is about exposing the framework through URIs - in effect providing RDF Lite. Also LINQ could provide neat scripting access to the store. Ok, it's far from certain they'll use e-r kit in revisions to IE, but they do have the tech on hand, and are promoting it heavily.

So while the current strategy may work in the short term for Moz, I suspect that a few years hence SQLite will be an albatross, and another deep revision of code will be needed to keep up with competitors. That may not be MS, I'd suggest that anyone that addresses the Web of Data directly will have a lead on anyone who doesn't.

Basically, given the steady growth in deployment of Semantic Web technologies, and developments like the LinkingOpenData project, the Moz folks should be careful that they're well-informed about what are "dark horses" rather than "lost causes". The Joost work should be a clue…

Meantime, I do hope Joost release their Redland material - above all, it sounds fun!

[btw Shelley, it's wonderful to be able to edit comments]

Thanks to all those who have contributed to the discussion. Comments are now closed, but you can contact the author of the post directly.

Prev: Guy Features?
Next: Traced Botanical Tour