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

Julian Reschke wrote:
> Sam Ruby wrote:
>> ...
>> Issue 53 is simply "Need to update media type registrations".  Outside 
>> of Larry, this part seems uncontroversial.  As to Larry, it simply 
>> seems to me a matter of question asked and answered[1].  Larry is 
>> welcome to correct me if I am misunderstanding.
>>
>> There are some technical details[2], but it seems to me that Ian is 
>> willing to accept bug reports on these.  The issue could be reopened 
>> should it turn out that Ian marks one or more of these bugs as 
>> WONTFIX, but we only would need to deal with that if it actually 
>> occurred and everything I see indicates that it will not.
>>
>> So, all that remains is the work.  From what I can see in section 13, 
>> the work has been done.  Other specs do it this way (in fact, I was 
>> associated with Atom, which is in the IETF and it did it this way -- 
>> see section 7 of RFC 4287), and it doesn't seem to me be a 
>> particularly egregious approach, so frankly I'm struggling to try to 
>> figure out what the real issue is here.
> 
> I agree that this is the common approach, and I have done that in other 
> RFCs as well.
> 
> However, I think the situation is a bit different for HTML, as the media 
> types are used for a variety of languages, and HTML5 does not describe 
> them all.
> 
>> Julian, you indicated that the IETF recommended an approach, and you 
>> were willing to do the work.  I fully appreciate that, but apparently 
>> that recommendation was made without the knowledge that the work is 
>> already done.
> 
> 1) The IETF AFAIK hasn't made a recommendation; and I don't think it can 
> or will. Change control for these media types is in the W3C.
> 
> 2) I made the recommendation *because* the editor added the registration 
> and I feel that this is not the best way to do it.

OK, so lets open bug reports or issues on what the Working Draft 
currently states.

>> I guess what I am asking here is:
>>
>>   1) Is there some significant technical issue with the approach
>>      that has been taken, or is this simply a matter of bug reports
>>      that have yet to be written?
> 
> The significant issue (IMHO) is that the media types apply to all HTML 
> languages, and the current RFC (2854) describing those takes these into 
> account. HTML5 does not.

Just so that I'm clear, from your perspective the core issue isn't "Need 
to update media type registrations", but rather that you believe that 
HTML 4.01 (as a concrete example) is a separate and distinct language 
from HTML 5.  If I am understanding correctly, I believe that's 
fundamentally different issue than than the issue that was raised.  If 
so, I would prefer that this issue be closed and a separate issue be raised.

[I feel compelled to disclose that I favor an approach of viewing HTML5 
as superseding prior versions of HTML, and treating as bugs any places 
where it doesn't adequately do so; but this in no way precludes opening 
of new issues]

>>   2) Is there reason to believe that the IESG would not accept the
>>      work that has already been done?  If not, I am having trouble
>>      understanding why ripping out this work and starting over
>>      would be simpler.
> 
> There is an existing document describing the media types, so the 
> simplest approach would be just to update it.
> 
> The editor has decided not to do that, but to in-line the registration. 
> This will work as well, as long as we do not lose significant information.
> 
>> If there are no technical issues (modulo a few bug reports, and from 
>> what I can see, there is an agreement that these are bugs), and what 
>> currently exists is workable, I'm inclined to recommend closing this 
>> issue.
>> ...
> 
> I'm not sure what bugs you refer to. The main issue is that if the media 
> type registration moves from RFC 2854 to HTML5, it should preserve 
> information about earlier versions of the language -- see 
> <http://tools.ietf.org/html/rfc2854#section-1>.

The bugs I was referring to was the comments that you made that Ian has 
missed a few places, specifically meta/@scheme, and head/@profile.  Ian 
indicated that he would get to them.  I'm just suggesting that if this 
is important to you, perhaps bugzilla would be a reasonable way to track 
the changes needed.

> BR, Julian

- Sam Ruby

Received on Monday, 24 August 2009 12:39:40 UTC