Re: ISSUE-41: extensibility change proposal advocates

Paul Cotton, Wed, 4 Aug 2010 17:20:11 +0000:

> Before the Chairs issue a survey on this issue we would like to 
> ensure that there is AT LEAST one "advocate" in the WG for each of 
> the five change proposals.
  ....
> Please reply to this email if you are an advocate FOR one or more of 
> the following change proposals with a clear indication of which 
> proposal(s) you support:
>
> 1. Change Proposal: "Proposal X", Registered XML-Style Namespace Prefixes 
> http://www.w3.org/html/wg/wiki/ChangeProposals/fixedprefixlikexml 

I like very that all of Robert Ennals' proposals operate with the 
concept "Valid extended HTML".

> 3. Change Proposal: XML-style namespaces with some naming restrictions. 
> http://www.w3.org/html/wg/wiki/ChangeProposals/html:xmlns 

I am prepared to advocate for the 3rd change proposal, "XML-style 
namespaces with some naming restrictions". I need to work more on my 
proposal, though, if time allows. 

My proposal pays special attention to the script element. Namespacing 
of the script element should be permitted, whenever SCRIPT also has a 
type attribute. This should be rather uncontroversial. There are 
several changes to the SCRIPT element in HTML5 compared to HTML4 
(escaping of the "/" character is not needed) and XHTML1.x (CDATA 
declaration is not needed). Thus <script> has become natural extension 
point. 

Robert, in the proposal mentioned below, suggests that text/html user 
agents only "activagtes" teh namespace declaration whenever his new 
@extension attribute is present. However, eventually, to use the type 
attribute seems more logical. Fore @type, there is already a MIME 
registration system - and thus we do not need to establish a namespace 
registry! Here I am inspired by Ample SDK. Consider what it does with 
the script element in this example: [1]

<script type="application/ample+xml"
    xmlns="http://www.w3.org/1999/xhtml"
    xmlns:aml="http://www.amplesdk.com/ns/aml">

Since I wrote my proposal, the @profile attribute has been 
(re)introduced on the scene. The profile attribute, as defined in RDFa, 
allows meaning to be attributed to prefixes without the use of 
namespaces. [2] This probably works fine for attribute values (which is 
where RDFa 1.1 uses them.) But thus I am also not sure that the profile 
attribute has anything to offer when it comes to prefixed attributes 
themselves. But I should review what my proposal says about prefixes, 
to see if needs updating or simplification.

> 4. Change Proposal: Generalize the mechanism used for SVG and MathML 
> http://www.w3.org/html/wg/wiki/ChangeProposals/extensionslikesvg

The analogy to how SVG/MathML has been incorporated is good - it takes 
care of the "what if it becomes standard in HTML".   However, I wonder 
if the @extension attribute that it proposes, eventually could be 
replaced with  type="application/langaugename+xml"? My own proposal 
suggests something similar for script, already - see above. Example: 
<foreignRoot type="image/foreignRoot+xml" xmlns="http://foo.ba/r" [...] 
>

The proposal doesn't have a solution for (real) attribute prefixes. 
What its says about RDFa is not quite correct: RDFa attributes does not 
not need namespace support, as the attributes of RDFa are in the XHTML 
namespace. It is the prefix used inside the _content_  of the RDFa 
attributes that rely on namespaces (in RDFa 1.0, but not in RDFa 1.1).

> If someone else advocates for a proposal you support there is NO need 
> to respond since we are looking for AT LEAST one "advocate".  This is 
> NOT an attempt to poll the WG on which change proposal is the most 
> popular.
> 
> This request for input closes on Fri Aug 13 at midnight ET.

[1] http://www.amplesdk.com/examples/extensions/
[2] http://ivan-herman.name/2010/08/03/new-rdfa-1-1-drafts/
-- 
leif halvard silli

Received on Thursday, 5 August 2010 03:23:32 UTC