On Tue, Sep 18, 2012 at 5:21 AM, Jake Archibald <jaffathecake@gmail.com>wrote: > On 18 September 2012 12:25, Robin Berjon <robin@w3.org> wrote: > > The more I think about this and read the (justified) use cases, the more > I > > tend to be convinced that it would be best to start with the simplest, > > smallest low-level approach that makes it possible to implement richer > > behaviour on top of. It's not a terribly original thought, but I would > > rather leave the intricacies involved in the sort of dependency > resolution > > you describe to library code. If common patterns emerge from that, we can > > then standardise those. > > The thing I like about the proxy solution is it's more (likely to be) > explicit, and as difficult as it looks. Whereas the manifest makes > things look simpler than they are, leading to 'gotchas'. > > It'd be interesting to see someone have a stab at a higher level > solution (will spend some time reading DataCache). I'm worried about > it going so low-level it becomes a barrier for most developers, like > WebGL. > The complexity of DataCache is daunting, to me; I'll spend more time digging in to it also, though. I did want to point to Chrome's INTERCEPT section ( http://code.google.com/p/chromium/issues/detail?id=101565) again, particularly the vague proposal for executable INTERCEPT handlers ( http://code.google.com/p/chromium/issues/detail?id=101800), which has not been implemented (Michael can correct me, but I don't think this is on the near-term implementation plan). Those ideas are all interrelated with DataCache, I think. -Chris
Received on Tuesday, 18 September 2012 18:36:54 UTC