Editorship Model for HTML and related specs Was: Please comment: Modular HTML or monolith?

> On Apr 6, 2016, at 08:11, Chaals McCathie Nevile <chaals@yandex-team.ru> wrote:
> 
> Hi all,
> 
> TL;DR: Should we move the content of specs like Shadow DOM and Custom Elements into HTML, or continue with the goal of more modular specifications?

(CCing / moving this to the AC Forum because I am broadening the topic to charter discussion)

I support modularity, especially if done in a way similar to what the CSSWG is doing: one central monolithic bedrock spec (HTML5.x / CSS2.1) and other specs that build upon it, either adding new features, or gradually refining/replacing sections of the bedrock spec.

However, regardless of my preferences, I still fail to understand how any model of editorship for HTML and related specs that is not in sync between the Web Platform WG and the WHATWG can fulfill this sentence from the Web Platform WG's charter:

> the Web Platform Working Group should make an effort to work with the WHATWG editors to avoid differences between the WHATWG and Web Platform WG specifications that would harm interoperability on the Web.

Personally, I prefer the model of governance, IP policy, and consensus based approach of W3C Working Groups over the way the WHATWG operates. However, as the Web Platform WG encourages and facilitates original work on its branch of specifications which are also maintained at the WHATWG, it is making things difficult.

Any would-be contributor to any of these specs has to decide for themselves which side they are supposed to contribute to. If I file a bug against the WHATWG's spec(s), and it gets fixed there, will it make it to the W3C's spec(s)? How about the other way around? How about new features? As people make this decision and individually pick one or the other, the specs are bound to diverge, and to do so in a way that is hard for anyone to track. This has always been a problem to some degree, but the sentence I quoted from the charter calls for attempting to minimize this problem, and until recently, it seemed somewhat possible to consider the WebApps / WebPlatform WG's specs as being downstream from the WHATWG's, and to act accordingly. However, the recent VCS changes, tooling changes, as well modularity discussions such as the one I am replying to, all make it easier to contribute new things to the W3C side. While this would be good in isolation, it is enacting the fork, and making this problem worse.

This seems to either be lending extra credence to this sentence from the WHATWG FAQ (emphasis mine), or an attempt at making the opposite case.

> The W3C publishes some forked versions of these specifications. We have requested that they stop publishing these but they have refused. They copy most of our fixes into their forks, but their forks are usually weeks to months behind. They also make intentional changes, and sometimes even unintentional changes, to their versions. **We highly recommend not paying any attention to the W3C forks of WHATWG standards.**

Either way, this does not sound like "making an effort [...] to avoid differences between the [...]" specifications". Are we operating on the assumption that implementors will ignore the WHATWG specs? If not (and I see no reason to believe this assumption is valid), we risk a colossal waste of resources, either by the W3C and its membership if we end up building a set of specs that nobody reads, or by implementors (and other spec readers), each individually trying to reconcile and make sense of both set of specs if they do attempt to work with both.

Although I did raise concerns at the time, I now regret not raising a formal objection to the Web Platform WG's charter over the fact that it does not specify for what purpose the WG works on specs that are also worked on at the WHATWG, and based on that purpose how it intends to operate. To be clear, I am not suggesting that there is no valid purpose for such an activity. Rather, I believe that there are several **conflicting** purposes that we could want to do this for, some of which I agree with, and others not. Unless we agree on which one we're pursuing, we have no basis for evaluating the best way to operate, or whether the WG is fulfilling its goals.

Maybe there is agreement between the chairs, or the editors, or w3m or some combination thereof, about what this purpose is. I do not believe however that there is such a shared understanding and agreement between the broader membership of the WG, let alone understanding by the public at large.

 - Florian

Received on Wednesday, 6 April 2016 03:05:03 UTC