Re: Amazon Silk

On 29/09/2011 18:48, "Yves Lafon" <ylafon@w3.org> wrote:
> On Thu, 29 Sep 2011, Robin Berjon wrote:
> 
>> So if these are indeed needs, then maybe we should make it possible (and
>> easier) for people to choose who they trust to accelerate their
>> browsing, as opposed to having a platform on which you can get fast
>> browsing from the tied-in browser, and a dog slow experience from all
>> the others. (Note that this isn't just a Silk issue ? Google's services
>> use SPDY with Chrome to make it much faster, something which people
>> would be screaming bloody murder about had say Microsoft tried it).
> 
> Well SPDY from chrome to Google's services is one thing, encrypting what
> goes from a browser to a content-transformation proxy is way different, as
> you have no real way to figure out if there is any abuse. And yes,
> perception by the public is one factor in the level of grudge that such
> things can raise.
> 

Interesting side-note: it looks like Amazon Silk is using SPDY in some
capacity:

http://aws.amazon.com/amazonsilk-jobs/

But to the general point of do we care about "split browsers" like Silk and
Mini for the purposes of the "linking" document: my view (discussed and
provisionally agreed with Jeni today) is that Web Architecture should treat
these as a single user agent since they are tightly coupled. This was the
logic we used when putting Opera Mini and things like it out of scope in the
Transforming Proxies best practices document in the Mobile Web Best
Practices WG.

> 
>>> In an alternate Universe, the W3C would have more control over the
>>> brand that is "the Web" (e.g., with a certification program, which is a
>>> notoriously difficult place for a standards body to go). I'm not sure
>>> if it would be a better or worse universe.
>> 
>> I honestly can't think of a single case in which such an approach
>> actually went well; but maybe it's just because we only hear of the
>> failures.

I dunno - Wifi Alliance? USB? UNIX?
...just sayin'...

Dan