[whatwg] A plea to Hixie to adopt <main>

[Tantek Çelik]

This was meant to follow-up to Henri's message[1]:

[1] http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-December/038219.html

but for some reason that didn't make it to my archives so I'm replying
to the latest message on this thread that did.


tl;dr: Having previously opposed the addition of a main element, I've
been convinced by the arguments (and counter-counter-arguments) and
evidence presented[1] that it's worth adding <main> to HTML.


I'm a strong advocate of being conservative of adding new elements to
HTML. Every element we add has a cost (in maintenance, learning etc.)
and perhaps increasingly so. That's the "high bar" that has been
referenced that has to be met - a new element must provide advantages
outweighing the cost to all of us of adding a new element. In
particular I am thinking of the cost to authors/developers of
continuing to grow HTML and its complexity.

<aside>
If anything I think I've grown more conservative regarding new
elements in this regard based on experience teaching authors. I used
to support <hgroup>, and though while I personally find it useful in
content, I no longer find its addition useful enough for authors in
general to overcome the confusion it adds. Similarly with <section>
(which appears to be turning into an alias for <div>). IMO the outline
algorithm is dead and we could simplify HTML by dropping these two.
</aside>

There has been a lot of reference to previous threads (on this list,
other lists, etc.) and at some point it becomes useless to say "go
search the mail archives" because no one has time follow all the
meandering threads.

I've written up a wiki page documenting what I believe to be
sufficient arguments to add the <main> element, along with arguments
against that I've heard and rebuttals, as well as counter-proposals
made along with flaws in counter-proposals:

http://wiki.whatwg.org/wiki/Main_element

Contributions/corrections/citations welcome, both for *and* against <main>.

>From discussions I've seen there appears to be a growing implementer
consensus that adding a <main> element helps more than it hurts and
thus I expect to see it happen.

However, I still think adding main is fully supported on principle
(rather than just on browser-implementation-Hixie-veto-override) and
thus I'm interested in capturing that on the wiki page so that
hopefully we can learn from this analysis about adding a new element
and use those lessons when considering new elements in the future.

Thanks,

Tantek