Re: [Webtransport] WGLC for draft-ietf-masque-h3-datagram

[Mark Nottingham <[e-mail]>]

Hi Chris et al,

I've had a read-through of the document, thanks.

The first thing that I noticed is that the specification is very obviously written from the perspective of someone who's very deep into the details of HTTP/3 and QUIC, and is either confusing or silent about how this relates to HTTP as a protocol overall. If this is going to be a new, version-independent feature of HTTP, I think we should specify it as one in the first instance -- especially given how poorly past attempts have failed when they weren't well-integrated (e.g., push).

So, I think it needs a non-trivial rewrite that shouldn't affect implementations, but unfortunately will affect the editors. I'm willing to work on a PR if that'd be helpful -- but it may take a bit of time to get right. Would that work for you/them?

Once that happens, I think it'll be easier to evaluate the technical content. If I understand it correctly, I have no problem with the on-the-wire details, although the capsule protocol feels like premature abstraction at this point. Could someone speak to the thinking behind it?

Cheers,


> On 22 Mar 2022, at 4:38 am, Christopher Wood <[e-mail]> wrote:
> 
> (Cross-posting to MASQUE, HTTPBIS, and WebTransport)
> 
> This email initiates the WGLC for draft-ietf-masque-h3-datagram, located here:
> 
>   https://datatracker.ietf.org/doc/draft-ietf-masque-h3-datagram/
> 
> Please review the document and send any comments to the MASQUE mailing list. HTTPBIS and WebTransport are cc'd given the overlap in technology.
> 
> This call will conclude on April 8.
> 
> Best,
> Chris and Eric

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