W3C home > Mailing lists > Public > public-html@w3.org > March 2018

Re: HTML 5.x, why?

From: Chaals Nevile <chaals@yandex.ru>
Date: Fri, 02 Mar 2018 11:53:13 +0100
Cc: "W3C Public HTML" <public-html@w3.org>
To: Léonie Watson <tink@tink.uk>, "Jens Oliver Meiert" <jens@meiert.com>
Message-ID: <op.ze8wizeknd6f5a@ordhord.home>
On Sun, 25 Feb 2018 19:52:07 +0100, Jens Oliver Meiert <jens@meiert.com>

> ...
> The strongest aspect to me seems to be the policy standpoint; and yet
> the question to me is whether that’s nearly as important for HTML as
> it is for WCAG. Overall the aftertaste remains that political reasons
> prevail.

Weighing in rather later...

I think politics tends to be pretty powerful, especially where there is a
sense that people have a right to make decisions for themselves. Anyway
there is certainly an element of politics here, and while supporting
things like different policy frameworks and different communities'
decisions on what matters to them is important to me, I don't think the
politics is all positive, reasonable and helpful. But that isn't what
determines whether it goes away, either.

> I don’t know whether that can be commented on here, let alone
> have it resolved, but the situation still seems unsatisfying :/

(I'm a co-chair, and I am commenting on it - I think it is a legitimate
aspect to discuss, given a respectful and thoughtful discussion.)

But there are other factors at play, including some that I think are
important to the people who are working on HTML at W3C, or relying on that
work. (That's the TL;DR: the rest of this just expands on that).

There are straight technical issues that have gone one way at WHATWG and
another way at W3C. Often, this comes down to how people interpret data -
because while data may hold the answers, the ones you get depend on the
questions you ask.

Although WHATWG began with the idea of documenting "what is, not what
should be", until recently WHATWG had been happily adding stuff hoping it
would become real, while W3C had moved to being much stricter about
matching reality.

Neither spec is perfect - or even close - but there is actually a "pretty
small" (yes, that is an entirely subjective claim) set of differences.

Some of the differences are deliberate: custom elements not being included
in e.g. 5.1 because it didn't have stable interop, 5.2 main and accesskey
matching user expectation and implementation reality respectively, and so
on. These differences are ultimately political in the sense that different
groups make different decisions, and technical in the sense that those
decisions lead to different understanding of the world, and in particular

I think it is easy to see these differences as more important than they
are. Many are about semantics of content, or things that still have fairly
patchy implementation (and a very few are just straight politics). The web
isn't a homogeneous monolith developing in lock-step, and people think
about what they want of bleeding edge functionality and really long-tail
compatibility, choose their sources of information (advice from friends or
experts, specs, tutorials, toolchains, etc) accordingly, and get on with
whatever they are actually trying to *do* on the web.

Some of the differences are known and slated to vanish as part of
producing HTML 5.3 - a few oddities deep in navigation, the aforementioned
custom elements.

There are also things that are unknown differences - plain bugs. This
category is not, to my mind, actually all that big, nor all that hard to
deal with, although it is a continuous (and often fairly boring) task so
long as HTML doesn't just stagnate with nothing new happening.

I understand the dissatisfaction. I also understand the dissatisfaction
of noticing that some corner of HTML is pretty bizarre, and the reason it
won't be changed is explained as not breaking the Web - which in one sense
is equivalent to "because we always did it that way". (There is more than
just that, for what it's worth, but that doesn't make the analogy
ridiculous :S).

In principle, as in at first, I like the idea of one source spec, one set
of documentation, one incubator, one authoring environment, one browser.
Too much mess isn't helpful. But monoculture has scary implications too.

Using Opera's mail client: http://www.opera.com/mail/
Received on Friday, 2 March 2018 10:54:02 UTC

This archive was generated by hypermail 2.3.1 : Friday, 2 March 2018 10:54:02 UTC