Re: HTML 5 and XHTML 2 combined

On 10/1/09 10:17, Philip TAYLOR wrote:
> Benjamin Hawkes-Lewis wrote:
>
>> 1. Processing is defined: the 'click' handler is executed first; then,
>> if the default action is not cancelled, the associated form (if any)
>> is submitted. The order is not defined for a element with two
>> _default_ functions.
>
> But that is just a lacuna in the current specification.
> There is absolutely no reason why the order of elaboration
> of two default behaviours should not be both specifiable
> and specified.
>
> (Which, in passing, addresses your earlier query that
> I deferred answering [snip]

Partially. My question is actually _how_ should it be specified?

For even more fun, how should a UA provide mechanisms for:

<submit submission="submit" href="http://en.wikipedia.org/wiki/HTML" 
cite="http://en.wikipedia.org/wiki/Fun"><label>Go!</label></submit>

or what happens you combine XHTML 2 with another XML language, like XLink:

<submit xlink:type="simple" 
xlink:href="http://en.wikipedia.org/wiki/Hyperlink" submission="submit" 
href="http://en.wikipedia.org/wiki/HTML" 
cite="http://en.wikipedia.org/wiki/Fun"><label>Go!</label></submit>

Specifying precedence of a set of DOM attributes - especially when some 
of the attributes might be imported from another language - seems a lot 
harder than specifying precedence based on DOM nesting, because the 
later has an intrinsic direction (parent to child or child to parent). 
Even if specified, it seems more likely to suffer from author error. 
Which brings us back to the simplicity of A HREF.

Is having multiple default functions activated by the same user action 
(clicking the submit button, say) even compatible with the spirit of 
providing a mechanism to activate those functions? Or should UAs be 
disaggregating the functions (perhaps presenting a menu on click and 
submenus for the options on right click?) Would the result be a good 
user experience?

In any case, how would activation of these functions work with 
DOMActivate events?

I'm not of course saying the spec needs to specify a particular user 
interface. But I think clarity about whether mechanisms should be 
aggregated or disaggregated, and order of functionality if aggregated, 
is crucial. And some examples would certainly help both content and user 
agent developers to imagine user experiences based on such features. I 
remember discussing the HTML4 "cite" attribute with some Elinks 
developers once, whose main problem was that the spec didn't give them 
an example of a mechanism to implement.

It seems to me there's a lot of complexity here which the spec doesn't 
try to untangle. I fundamentally agree it could try to do so, but it's 
not clear to me that the feature's so very advantageous in the first place.

--
Benjamin Hawkes-Lewis

Received on Saturday, 10 January 2009 11:10:08 UTC