Re: W3C writes spec for distributed tracing for HTTP APIs

I am one of the people who has asserted something like this--although if my
comments are part of Manu's allusion, I would frame them a bit differently.
I think the W3C regularly works on conventions that put HTTP to work in
useful ways; what I see less of is the W3C successfully promoting standards
that deal with any other transport, or with any protocol that is not
layered on HTTP. It is, after all, a standards org for the web, as its name
implies. All of the examples Manu cites accord with that characterization;
even the thing description spec, which mentions other protocol bindings in
section 8.3.2, gives no concrete instantiations other than HTTP. I would
also note that although specs exist, this is not evidence that they have
traction.

This is not to take serious issue with the specifics in Manu's comment,
which is indeed a good data point. I only to suggest a different framing. I
think the interesting questions are: A) is a spec inherently web-centric
(should its foundation be HTTP)? B) Will the W3C be a good source of
momentum for it? C) Is a credentials community group the right home, within
the larger W3C umbrella? Given a yes answer to all three of those
questions, incubating at W3C CCG makes perfect sense. But I think many of
the topics we are discussing don't generate a yes answer to all three of
these questions.

On Fri, Jan 3, 2020 at 9:34 AM Manu Sporny <msporny@digitalbazaar.com>
wrote:

> Recently, a number of people note that W3C doesn't do protocol/HTTP
> APIs, even though there are examples at W3C to the contrary.
>
> This W3C proposed standard (the very last step before global standard)
> just came across my inbox and is an interesting data point that argues
> against the premise that "W3C doesn't work on [successful] protocol/HTTP
> APIs":
>
> https://www.w3.org/TR/trace-context/
>
> The specification is particularly interesting because it 1) is a
> distributed systems specification, 2) is a collaboration between
> Microsoft and Google, 3) focuses on an HTTP protocol/API, and 4)
> reinforces the notion that W3C *does* work on this sort of stuff.
>
> Here are a few other W3C specifications that define protocols, HTTP
> APIs, and distributed systems:
>
> https://www.w3.org/TR/wot-thing-description/#http-binding-assertions
> https://www.w3.org/TR/webrtc/
> https://www.w3.org/TR/activitypub/
> https://www.w3.org/TR/ldp/
>
> It's one of a number of data points that folks should be aware of when
> contemplating where to do future protocol / HTTP API work that this
> group (and others) incubate.
>
> -- manu
>
> --
> Manu Sporny (skype: msporny, twitter: manusporny)
> Founder/CEO - Digital Bazaar, Inc.
> blog: Veres One Decentralized Identifier Blockchain Launches
> https://tinyurl.com/veres-one-launches
>
>