Notes from Edge Conf

Like many others, I found myself signing a document guaranteeing Facebook’s privacy as I wandered into their swanky offices in Covent Garden that for the day-long mega geekout Edge, a “conference on advanced web technologies for developers and browser vendors” organised by Andrew Betts and his chums at FT Labs. And his dog Shadow.

I was unsure how the format would work, as it was all panels. My experience of panels at SxSW is that they are the only way to pleasure your best friends around a desk in public without breaking the law. It worked very well, though, due to good moderation, good speakers and a highly clueful engaged audience.

The videos are going online courtesy of Pete Le Page.

Some nuggets I got from the event follow in random order.

Everyone connected to the conference is hiring: FT Labs, Facebook, Google. Most of Google were at the event. I assume they’re not all allowed to travel in the same plane, or there would be no-one in the world who understands Web Components if something went wrong. No-one from Apple or Microsoft appeared to be there.

Appcache will soon (please please please) be superseded by something called Application Controller. Here’s an example of how it will be able to deal with responsive images. See Jake’s presentation (the first 15 minutes of the offline panel).

Forcing CSS transforms to be 3D in order to get them hardware accelerated can misfire on mobile, making it less performant due to time it takes to move from CPU to GPU.

Everyone seems to agree that different vendor-proprietary manifest formats for packaged applications is bad and will inhibit uptake. No-one expressed any willingness to standardise it, though. Similarly with device APIs. Why not just all adopt the syntax that PhoneGap uses? That’s what devs are familiar with now.

Microsoft’s Pointer Events API submission to W3C is good, say Chrome people. Opera people like it too. No word from Mozilla and Apple seemed somewhat hostile. Boris Smus (I believe) was insightful about not adding a z axis for 3d gestures because it removes the mapping to the device’s input surface: what would the z axis units be?

From Remy: if you’re using Web Sockets over mobile, you must use SSL, otherwise network providers are likely to block it.

I didn’t hear any discussion of DRM (although there was plenty in the bar the night before.)

It was an excellent day, full of high quality information. Congratulations to Andrew Betts, his FT Labs colleagues and Facebook friends for organising it.

Comments are closed.