Rechartering HTTPbis

Everyone,

We're getting close to Working Group Last Call. There are about 10 or so outstanding design issues, and I'd like them to be closed by Paris as we're now past the four year mark (on a WG that was originally chartered for a year and a half).

When this effort was started we took great pains to make it clear that we weren't working on a new version of HTTP, because there wasn't implementer support to do so and we wanted to focus upon interoperability and security.

That's clearly changed in the intervening time; two major browsers have implemented SPDY, a non-textual serialisation of HTTP's semantics, and there are now several other implementations as well (full disclosure: including an experimental one in Python by yours truly). 

I've been talking to a number of folks -- including those implementing SPDY, as well as HTTP implementers and the W3C TAG -- about this recently. There seems to be broad agreement that the time is ripe to start work on a new version of HTTP in the IETF, and that it should happen in this Working Group.

Why here? This mailing list is the best approximation of the HTTP community; it has participation (or at least presence) from most implementations, including browsers, servers, intermediaries, CDNs, libraries, tools, etc. I firmly believe that as HTTP evolves, it needs to accommodate the entire community, not just the selected needs of a subset, so rather than creating a new WG or having a private collaboration, it should happen here. 

I've put together a charter proposal (see attached) that has us going to WGLC shortly (something that I want to see us do regardless), and starting work on HTTP/2.0. Note that it does NOT call out a starting point; rather, we'll start by asking for proposals, considering them and selecting one based upon the traditional IETF criteria of rough consensus and running code.

We'll then spend about a year refining that proposal to make sure it is a suitable evolution path for HTTP, while offering better performance, security and interoperability.

Please have a look and tell us your thoughts, indicate support, or express any concern you might have. I'm also happy to talk to you privately if preferable. If there's good support in the WG for doing this, I plan on taking it to the IESG before Paris, so that we can have two meeting slots; one to work on BIS issues (hopefully, WGLC ones), and one to discuss HTTP/2.0.

Regards,

--
Mark Nottingham
http://www.mnot.net/