W3C home > Mailing lists > Public > public-html@w3.org > August 2009

Re: ISSUE-53: mediatypereg - suggest closing on 2009-09-03

From: Roy T. Fielding <fielding@gbiv.com>
Date: Tue, 25 Aug 2009 15:27:00 -0700
Message-Id: <555D7D73-DB53-4E05-8715-EB22AA9528A7@gbiv.com>
Cc: "public-html@w3.org WG" <public-html@w3.org>
To: Maciej Stachowiak <mjs@apple.com>
On Aug 24, 2009, at 11:42 PM, Maciej Stachowiak wrote:
>> I don't think that section 12.2 satisfies the issue.  What is wrong
>> with defining the elements and attributes where an implementor of
>> "text/html" is going to need to know about them and simply mark those
>> features as deprecated?  For example, <a name> has required  
>> processing
>> associated with it, so why not just define that under <a>?
> As an implementor, I think the required processing is sufficiently  
> defined,

As an implementor of content management systems, HTML generators,
and link management tools, I do not consider it defined at all.

The issue is regarding the definition of "text/html".  The purpose
of media types is to allow the sender to tell the recipient how
the data should be interpreted.  In some cases, that recipient is
a browser.  A recipient is supposed to be able to follow its nose
from the media type to the registration template for that media
type to the specification(s) listed as definitive for that media
type in order to find out what the data within that message
actually means.

Data marked as "text/html" contains <a name> attributes.
It has a specific meaning as mark-up in the language.  That
meaning can either be defined by listing all of the HTML
specifications as definitive for "text/html" or by defining
all of the text/html syntax within the one specification of HTML.
So, before this issue can be marked as resolved or closed,
one of the above two options must be reflected in the
state of the draft specification.

> and in the places where I would look when implementing, testing or  
> refining the pieces of code where its behavior would be exposed.  
> The name is attribute is mainly relevant to implementors not when  
> implementing the element, but rather when implementing the various  
> collections and other APIs that expose certain elements accor

Maybe, if browsers are the only things in the world that process
HTML.  They are not.

> Do you think anything specific about <a name> (whether authoring  
> requirements, implementation requirements or anything else) needs  
> to be defined, but isn't? Or is it just a matter of where in the  
> spec the information is found? I can file a bug for you if you feel  
> that needed info is missing and if you can tell me what it is.


HTML 4.01 spec has attribute definitions.  HTML5 draft does not.


Copy and paste is the solution.

Received on Tuesday, 25 August 2009 22:27:32 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:50 UTC