W3C home > Mailing lists > Public > public-webcrypto@w3.org > January 2016

Re: [W3C Web Crypto WG] how to progress ?

From: Ryan Sleevi <sleevi@google.com>
Date: Thu, 21 Jan 2016 15:56:04 -0800
Message-ID: <CACvaWvaBhaWQ6_LAsK5L2eLapXDj1NosXWx7vqW_UP687EQoBA@mail.gmail.com>
To: GALINDO Virginie <Virginie.Galindo@gemalto.com>
Cc: "public-webcrypto@w3.org" <public-webcrypto@w3.org>
On Thu, Jan 21, 2016 at 8:57 AM, GALINDO Virginie <
Virginie.Galindo@gemalto.com> wrote:

> Dear all,
>
>
>
> After the conversation between Harry and Ryan, completed by the recent
> categorization of bugs done by Harry and clarified by Eric, I understand
> the following :
>
> -          Implementations of the web crypto seems to be still in debug
>
> -          UAs contributions to the web crypto WG does not seem to be
> high priority (Mozilla mentioning next months, Google mentioning reduced
> bandwidth)
>
>
>
> We have two paths to progress :
>
> -          One :  we try to deliver something quick, gathering the
> minimum core of interoperable features (but we will lose our editor who is
> not supportive to this option) (Ryan, I let you correct if I interpreted
> wrongly your emails). In that case we answer the developers expectations to
> have a spec.
>
> -          Two : we give us 5 to 6 months to let implementations being
> debugged, UA discussing interoperable features and align the Web Crypto API
> accordingly (we will have to find a way in W3C to adapt the spec timeline,
> super-limit-chartering aspects, but I am confident we will find a way). In
> that case we make developer waiting but deliver a wider range of features.
>
>
>
> I would like to hear from the WG members preferred **strategy** asap, so
> that we can analyze with W3C the best way to move forward.
>

Virginie,

I believe that regardless of the direction the WG takes (and I think Two is
what we'd be supportive of, but I don't think there'd be support for One at
all), I unfortunately need to step down as editor. Given far more pressing
responsibilities, I simply don't have the time, or, unfortunately, the
professional energy, to continue on as editor for the spec. That said, both
I and the Chrome team remain committed to progressing on WebCrypto in
implementation and within the working group, and are hoping to see other
UAs take a more proactive role in helping find an interoperable, robust
solution for users.

While I think the spec is largely in a good place, there's definitely signs
of interoperability issues and spec-compliance issues. While I think these
issues are entirely surmountable, I think any proposal which requires
excising entire features from the spec - as has been requested of KeyFormat
and as suggested in One - means there are significant editorial commitments
needed. Worse, if that's accepted as the standard for publication - minor
misalignment causing major removals - then I think we'd find a similar
large number of core features needing to be removed. The proposed
resolution of that - creating two versions of the spec, one of which
represents only the subset that is wholly implemented, and one of which
represents both aspirational and mostly-implemented features - is just way,
way too much work, and without other UAs to collaborate on this list and on
the spec, is simply not a worthwhile use of time. Worse, with just a small
amount of collaboration and work from other UAs, almost all of those
changes would be made unnecessary.

While I think Option Two is ideal, I don't know if it's realistically
achievable without greater involvement by members and UAs. If that's the
path the WG takes, we simply need greater involvement, as well as a focus
on 'spec compliance' and not just 'rough interoperability,' which I do
believe we have today. Do I think every UA needs to fully adhere to the
spec in all of its nuance for us to ship to PR? No, of course not. But I
think we at least need other UAs making commitments to either implement
what the spec says (treating it as browser bugs), or clarifying if they
object to the spec language or requirements, so that we can actively
resolve these issues. Do I think everything in the spec is either necessary
or correct? Probably not. But that's why feedback from other implementors
is needed - given the diversity of approaches and goals, that sort of
feedback is key. We need that consensus to be active - I don't think a
"silence is assent," as has been proposed and executed for KeyFormat,
really represents the best outcome, and seems rather arbitrary.

But regardless of these positions, as I mentioned, I think the time
commitment necessary for either position is, unfortunately, far greater
than I can honestly make, and as such, need to step down as editor. Either
option will require more hours-per-week of editorial work than I'm able to
commit to, and I would rather not be the one holding the WG back from
progressing. :)
Received on Thursday, 21 January 2016 23:57:13 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 21 January 2016 23:57:14 UTC