Dan Burnett | 1 Aug 22:17 2011

questions: partial interfaces, text for WebIDL Constructor

Hi.  I'm a new user of ReSpec.js, having started with it last week as the framework for the WebRTC
specification.  I've run into a couple of WebIDL items I can't seem to figure out and wondered if anyone on
this list might be able to help.  They are:

1) Specifying a "partial interface" seems to result in a failure.  In particular, specifying

partial interface URL {
  static DOMString createObjectURL(in MediaStream stream);
};

as 

<dl title='partial interface URL' class='idl'>
  <dt>static DOMString createObjectURL (in MediaStream stream)</dt>
  <dd>Text description here</dd>
</dl>

gives me the following error:

	• Expected definition, got: partial interface URL
	• Processing error: [Exception... "Component returned failure code: 0x80004003
(NS_ERROR_INVALID_POINTER) [nsIDOMDocumentFragment.appendChild]" nsresult: "0x80004003
(NS_ERROR_INVALID_POINTER)" location: "JS frame ::
file:///Users/burn/Documents/webrtc/2009/dap/ReSpec.js/js/respec.js :: anonymous :: line 1976"
data: no]

Are partial interfaces not supported?

2)  I would like to provide a text description for the constructor method ("new Blah") itself, but I haven't
yet found a way.  Has anyone else figured out a way to get the constructor itself to show up in the list of
(Continue reading)

Aryeh Gregor | 18 Aug 21:59 2011

Publication of specifications as HTML5

Today I was talking with a few people about the fact that W3C
specifications cannot be published as HTML5.  As far as I can tell,
the pubrules say:

"All normative representations must validate as one of the following:
HTML 4.x, some version of XHTML that is a W3C Recommendation, or RDFa
in XHTML. Team Contacts please see the Communications Team to propose
additional exceptions."

HTML5 has been under development in some form for over seven years, at
the W3C for about four years, and is unlikely to reach Recommendation
for a number of years yet.  It introduces many new semantic elements
that would be useful in specifications, such as <nav>, <header>,
<section>, etc.  According to Ian Hickson, HTML3.2, HTML4, HTML4.01,
XHTML1, and XHTML1.1 were all allowed to be published in their own
format even before they reached Recommendation.  Hixie also said that
he reformats the HTML5 specification from HTML5 to HTML4.01 with a
four-line Perl script, which implies that if any consumer actually
needs HTML4.01 instead of HTML5, converting it should not be a great
burden.

So it's a bit of a puzzle to me why specs can't just be published as
HTML5.  What are the practical problems it might cause that outweigh
the benefits?  Multiple people have told me that polyglot HTML5 might
be okay, but why is non-polyglot HTML5 any worse than HTML4.01 (which
is allowed)?  What would the procedure be for trying to get this
requirement changed?

Thanks to everyone for their time.

(Continue reading)

Doug Schepers | 18 Aug 22:15 2011
Picon

Re: Publication of specifications as HTML5

Hey, folks-

On 8/18/11 3:59 PM, Aryeh Gregor wrote:
> Today I was talking with a few people about the fact that W3C
> specifications cannot be published as HTML5.  As far as I can tell,
> the pubrules say:
>
> "All normative representations must validate as one of the following:
> HTML 4.x, some version of XHTML that is a W3C Recommendation, or RDFa
> in XHTML. Team Contacts please see the Communications Team to propose
> additional exceptions."
...
> So it's a bit of a puzzle to me why specs can't just be published as
> HTML5.  What are the practical problems it might cause that outweigh
> the benefits?  Multiple people have told me that polyglot HTML5 might
> be okay, but why is non-polyglot HTML5 any worse than HTML4.01 (which
> is allowed)?  What would the procedure be for trying to get this
> requirement changed?
>
> Thanks to everyone for their time.

Speaking for myself, I would support using HTML5 in TR, whether polyglot 
or not.

For those that use an XML toolchain, couldn't they simply convert 
HTML5-based specs to their format of choice?

Regards-
-Doug

(Continue reading)

Karl Dubost | 18 Aug 22:25 2011
Picon

Re: Publication of specifications as HTML5


Le 18 août 2011 à 15:59, Aryeh Gregor a écrit :
> What are the practical problems 

I'm interested by this part of the discussion specifically.
Could we outline the issues that would have to be solved if we were injecting an HTML5 specification in the
current system?

Then from here, we could move on on helping Ian Jacobs (and current Webmaster) to adjust these tools so we
could meet halfway.

--

-- 
Karl Dubost
Montréal, QC, Canada
http://www.la-grange.net/karl/

Ian Jacobs | 18 Aug 23:59 2011
Picon

Re: Publication of specifications as HTML5


On 18 Aug 2011, at 3:25 PM, Karl Dubost wrote:

> 
> Le 18 août 2011 à 15:59, Aryeh Gregor a écrit :
>> What are the practical problems 
> 
> I'm interested by this part of the discussion specifically.
> Could we outline the issues that would have to be solved if we were injecting an HTML5 specification in the
current system?
> 
> Then from here, we could move on on helping Ian Jacobs (and current Webmaster) to adjust these tools so we
could meet halfway.
> 

Historically we have waited until a spec is further along the standards track before adding it to pubrules.
There has been a growing demand to use HTML5 for TRs. 

The staff has been discussing this. Karl mentions tooling issues (specifically the pubrules checker)
which is one of the topics we've been discussing. Another one has to do with polyglot support. We've not
completed our discussions internally, but discussion on spec-prod is also welcome.

I'll keep an eye on this thread and will have more to contribute once the team discussions have wrapped up
(slowed by vacations).

Ian

--
Ian Jacobs (ij <at> w3.org)    http://www.w3.org/People/Jacobs/
Tel:                                      +1 718 260 9447
(Continue reading)

Karl Dubost | 19 Aug 04:12 2011
Picon

Re: Publication of specifications as HTML5


Le 18 août 2011 à 17:59, Ian Jacobs a écrit :
> Another one has to do with polyglot support.

* What would be the benefits in the context of spec publishing to have polyglot support?
* What do you define as polyglot support?

--

-- 
Karl Dubost
Montréal, QC, Canada
http://www.la-grange.net/karl/

Ian Jacobs | 19 Aug 05:20 2011
Picon

Re: Publication of specifications as HTML5


On 18 Aug 2011, at 9:12 PM, Karl Dubost wrote:

> 
> Le 18 août 2011 à 17:59, Ian Jacobs a écrit :
>> Another one has to do with polyglot support.
> 
> * What would be the benefits in the context of spec publishing to have polyglot support?
> * What do you define as polyglot support?

I had understood "conforms to http://www.w3.org/TR/html-polyglot/"

For XML processors.

Ian

> 
> 
> -- 
> Karl Dubost
> Montréal, QC, Canada
> http://www.la-grange.net/karl/
> 
> 

--
Ian Jacobs (ij <at> w3.org)    http://www.w3.org/People/Jacobs/
Tel:                                      +1 718 260 9447

(Continue reading)

Karl Dubost | 19 Aug 05:28 2011
Picon

Re: Publication of specifications as HTML5


Le 18 août 2011 à 23:20, Ian Jacobs a écrit :
> For XML processors.

If all the tags are closed and the attributes are quoted (aka "XML well-formed"), would it be enough for the
XML pubrules tools?

--

-- 
Karl Dubost
Montréal, QC, Canada
http://www.la-grange.net/karl/

Ian Jacobs | 19 Aug 05:34 2011
Picon

Re: Publication of specifications as HTML5


On 18 Aug 2011, at 10:28 PM, Karl Dubost wrote:

> 
> Le 18 août 2011 à 23:20, Ian Jacobs a écrit :
>> For XML processors.
> 
> 
> If all the tags are closed and the attributes are quoted (aka "XML well-formed"), would it be enough for the
XML pubrules tools?

The original goal was not to make the pubrules checker happy, but to have a version readily consumable by any
xml consumer. It turns out that will help the existing pubrules checker, but that's not the main goal.

Ian

> 
> 
> -- 
> Karl Dubost
> Montréal, QC, Canada
> http://www.la-grange.net/karl/
> 
> 

--
Ian Jacobs (ij <at> w3.org)    http://www.w3.org/People/Jacobs/
Tel:                                      +1 718 260 9447

(Continue reading)

Doug Schepers | 19 Aug 06:59 2011
Picon

Re: Publication of specifications as HTML5

Hey, folks-

On 8/18/11 11:34 PM, Ian Jacobs wrote:
>
> On 18 Aug 2011, at 10:28 PM, Karl Dubost wrote:
>
>> Le 18 août 2011 à 23:20, Ian Jacobs a écrit :
>>> For XML processors.
>>
>> If all the tags are closed and the attributes are quoted (aka "XML
>> well-formed"), would it be enough for the XML pubrules tools?
>
> The original goal was not to make the pubrules checker happy, but to
> have a version readily consumable by any xml consumer. It turns out
> that will help the existing pubrules checker, but that's not the main
> goal.

Arguably, people using the XML toolchain would be in a better position 
to adapt the content to their needs than the average Web developer, who 
may benefit more from the simpler syntax of HTML5.

Regards-
-Doug


Gmane