Re: Decentralised extensibility idea (ISSUE-41)

On Fri, 15 Jan 2010 18:51:58 +0100, Leif Halvard Silli  
<xn--mlform-iua@målform.no> wrote:

> Philip Jägenstedt, Fri, 15 Jan 2010 18:05:09 +0100:
>> On Fri, 15 Jan 2010 17:13:43 +0100, Toby Inkster <tai@g5n.co.uk> wrote:
>>
>>> Philip Jägenstedt wrote:
>>>
>>>> I don't think this is a very good idea, as data-* are always hidden
>>>> and not suitable for marking up content that is visible in the page.
>>>
>>> You are mistaking my proposal for a method of embedding data into a
>>> document. The proposal is not intended for embedding the kind of data
>>> that Microdata or RDFa embed; rather it's a general purpose extension
>>> point that other standards ("otherspecs") could use.
>>
>> Yes, I did assume as much from the examples, but in fact I think they
>> are effectively the same thing.
>
> Same thing? How?

Because the only thing you can do unless you're a user agent vendor is add  
dumb data and use JavaScript/CSS/whatever to make your extension do  
something interesting. User agents on the other hand will simply add  
native HTML elements and attributes and then try to have them  
standardized. I don't see any problems left solving, honestly.

>> Only user agents can change the behavior of any element/attribute in
>> other ways than what is already possible using JavaScript and CSS. If
>> you are a user agent (especially a browser) then I would argue that
>> you *shouldn't* be making stuff up, you should make a
>> proof-of-concept and then propose the feature for standardization as
>> a proper HTML feature. Centralization here is a good thing, because
>> it makes vendors talk to each other and improve the feature before it
>> is too late.
>>
>> If, on the other hand, you are not a user agent, then the only thing
>> you can do is embed data and make any behavior/rendering with
>> JavaScript/CSS. For embedding the data/hooks you need you can use
>> data-* attributes, clasa attributes, microdata or whatever you want.
>
> Jonas has meant that D.E. could be handled by MD. As such, it would be
> possible to (mis)use MD as a "general purpose extension". By your logic
> then, we should not have MD, as it allows non-UA vendors to "avoid" the
> W3 centralisation.

Microdata is specifically for semantics which isn't and most likely won't  
be standardized in HTML because their use is not common enough. However,  
it is obviously a good thing to standardize very common vocabularies, be  
that at the W3C or somewhere else. Microdata can be used for "general  
purpose extension" if you want, but only because the only form of  
extension you can add unless you're a user agent is embedding data.

> It seems like the WG needs to decide whether at all there should be any
> way for non-UA vendors to decide/define how HTML should interpreted ...
> ? Aren't you simply not in tune with the charter here?

Non-UA vendors *cannot* change the processing of HTML, more or less by  
definition. Do quote the charter if you think it's relevant.

> Authors can use "data-* attributes, clasa attributes, microdata or
> whatever you want" - but not @profile? Give me a break. A powerful
> vendor like Google already use data-* for their "SVG-via-Flash"
> solution. It is just naïve to not realize that they are effectively
> defining a profile?

I'm not sure what you are arguing for or against - what has @profile got  
to do with anything? Things wouldn't be any different if the proposal was  
to use the @pineapple attribute. (I haven't followed the endless  
discussions on @profile and don't have any opinions on the attribute as  
such.)

-- 
Philip Jägenstedt
Core Developer
Opera Software

Received on Friday, 15 January 2010 20:07:10 UTC