Re: WHATWG/W3C collaboration proposal

On 11/25/2014 11:54 AM, Domenic Denicola wrote:
> In general this proposal seems good, and I'm very heartened by the
> responses here from various parties. However, it reminds me of
> something Robin and I tried with URL a couple months back that ended
> up floundering on irreconcilable differences, so I wanted to get
> those potential issues out in the open before we go too far down this
> path. I remain hopeful though!

Thanks.  Even though I disagree (see below) this is valuable.  I plan to 
continue to be up-front, open, and public with my plans.

> The WHATWG's primary concern in any scenario like this is that the
> snapshots produced do *not* look like or act like reference-able
> specifications meant to be implemented or consulted by authors. This
> is a general principle, and any good-faith efforts to collaborate
> need to work within that framework. For example, here are two issues
> which we were not able to resolve last time around:

You and I are both WHATWG spec editors, yet we disagree.  I'm curious as 
to what extent this qualifies a "WHATWG's primary concern".

I will also note that Anne is well aware of what I am doing, and has is 
OK with what I am doing.  At least two times I have pushed too hard or 
too fast, and each time I've circled back to ensure that Anne is OK with 
my proceeding.

> 1. The snapshot must have no/minimal stylesheet.
> is an example. 2. The
> snapshot must have a title, or subtitle, that clearly reflects the
> purposes of the snapshot, and that it should not be used for
> implementations. We want to avoid "snarky", but something like "For
> IPR purposes" or "Not for implementations" would be important. (For
> those interested in commit-level snapshots for reference purposes, we
> can get the WHATWG producing and hosting such snapshots very
> quickly---within a week, I would anticipate. See
> for previous discussion of this.)
> It's also important not to make references to the source WHATWG
> specification as "editor's drafts." The WHATWG is not a source for
> editor's drafts.
> If the plan is, as Sam says and Jeff supports, that the copies would
> be virtually byte-for-byte, this should be no problem. The WHATWG can
> produce such snapshots easily, more or less as was done for URL
> already, but with a more agreeable title. We can even address Jeff's
> concerns by baking the W3C copyright statements and boilerplate into
> the header, much like was done with the FSA boilerplate for
> (That way, the snapshots
> don't have to be modified upon copying.)
> Finally, Sam already emphasized this in his post, but I want to
> re-emphasize it: an arrangement like this works best if all involved
> try to avoid forks. This means changes are made in the source
> document, then flow upstream to the snapshot. As such, it would be
> appreciated if the collaborative feedback and community Sam is hoping
> to build was centered on the list and whatwg/url
> GitHub repo, instead of e.g. confined to public-webapps or to
> whatever GitHub repo the snapshot ends up in.
> So, what do you guys think? Is this feasible?

I don't agree with this.

If you look at Living Standards and/or recent editors drafts you will
generally find good stuff.  You will find things that are the right
direction, generally agreed to by all, but not widely implemented just
yet.  You will also find stuff that isn't quite baked yet, not widely
reviewed, and may change drastically.  And you will find flights of
fancy.  In some cases, you will even find stuff that implementers have
tried, found distasteful, and actively rejected but still remain in the
spec years later.

I can provide examples of each, but in the interest of not derailing the
conversation, I'll decline to do so.

The point being that there are people who value exactly what I cited in
my blog post:

    In particular, they are interested in evidence of wide review,
    evidence that issues have been addressed, evidence that there are
    implementations, and the IPR commitments that are captured along the

And I'll restate for emphasis a point I made in my blog post:

    The fact that there are people who donít doesnít make the first set
    of people go away or become any less important.

I will state that my plans are to meet those needs.  My plans are to 
also build a URL specification that is usable by both end users and 
developers alike.  And for that specification to be brutally honest 
about what features are implementation dependent and implementation 
defined.  And to actively work with all concerned to make those last two 
sets asymptotically approach zero over time.

- Sam Ruby

Received on Tuesday, 25 November 2014 17:24:07 UTC