Tracking of pending media type/charset/URI registrations

Looking over draft-iab-iana: I think this document entirely misses two things:

1) the widespread use of unregistered protocol parameters (or their use for meanings other than exist in the registry), and the resistance of protocol designers -- especially at the application level -- to participate in the IANA processes.

2) the desire to map registered values to stable values, or to use URI assignment rather than registry processes as a way of distributing the extensibility of protocols.

Maybe IAB doesn't see the same problems in other (non-application) registries?

Larrry




-----Original Message-----
From: Larry Masinter 
Sent: Sunday, February 20, 2011 4:39 PM
To: 'Graham Klyne'
Cc: Michelle Cotton; Alexey Melnikov; Philippe Le Hégaret; Roy Fielding; www-tag@w3.org
Subject: RE: Tracking of pending media type/charset/URI registrations

(Moving cc from tag to public www-tag; noting that this is part of my tag ACTION-531 "Write draft document on architectural good practice relating to registries" taken at the last TAG face-to-face meeting):

I believe that there are a number of voices  who would rather see MIME types, URI schemes, charset declarations, link relationships, and other values now defined as IANA registries instead be maintained in a Wiki (presumably one in which they and everyone else have some editorial control), or by some organization over which they feel like they have some control.

In some cases, groups have 'routed around' the registries by proposing protocols which reuse registered values with other meanings (using a "willful violation").  In at least one case, a  W3C working groups felt that establishing and managing an IANA registry  was inappropriate, and instead started a W3C registry.  

I don't think the issue is solely a "web" or "W3C" issue, though. I ran into a similar issue with 3GPP and the media feature registry when I was expert reviewer for those values (used in SIP), which is why I'd like to see if there was a more general IANA solution rather than a narrower W3C solution or an even narrower HTML solution. 

I think this is as much a political/power issue than an operational process one.  (I am reminded of the struggle for management of DNS root virtual real estate.)

The fact that the web is full of deployed products and services that happily use unregistered values for  things without IANA entries leads me to believe there is a problem which is broader than "transparency" or "education" or "miscommunication". 

I'm wonderin if there is a hybrid approach, where preliminary values can be documented in a Wiki or tracker or other mechanism, but that more open space is also linked to a more "official" registry in which registered values have undergone the review process originally designated for them, including expert review or community consensus. 

Larry
--
http://larry.masinter.net

-----Original Message-----
From: Graham Klyne [mailto:GK-lists@ninebynine.org] 
Sent: Friday, February 18, 2011 2:42 AM
To: Larry Masinter
Cc: Michelle Cotton; Alexey Melnikov; Philippe Le Hégaret; Roy Fielding; tag@w3.org
Subject: Re: Tracking of pending media type/charset/URI registrations

Yes, I think improving transparency and disclosure (and timeliness) would be 
positive.

I know I'm sometimes a bit slow in responding to registration review requests 
(on the order of weeks, not months), but I'm unaware of any case where a 
provisional scheme registration has been prevented from proceeding reasonably 
promptly.

There was a case last year when I asked the requesting group to delay a 
registration (ws:) because there was some live discussion about the choice of 
scheme name, but I think I made it clear the request wasn't being blocked.

I suspect, but have little evidence, that the perceived problems are a lack of 
awareness of the lightweight procedure available for provisional registrations 
(of URI schemes).

Maybe a practical approach might be for any request to result in an immediate 
"registration requested" or "pending" entry in the provisional registry, so that 
people can see a quick response.  This has a disadvantage that each request 
would result in additional IANA actions (one to post the request, then to remove 
the "pending" flag.

#g
--

Larry Masinter wrote:
> I don't think the problem is that IANA isn't doing the right thing,
> it's more that there is a gap between what IANA and IESG and expert
> reviewers are doing vs. what many in the community seem to want --
> a light-weight way of noting things that *should* be registered,
> an easy way of finding out about registration applications and 
> expert reviewer comments, etc.
> 
> I don't have the complete list of requirements and it may be premature
> to conclude what the solutions are, which is why I'd like to gather
> together.
> 
> Some of the difficulties may be rooted in a power struggle, over
> "who is in charge", partially analogous  to the issues that arose
> over control of the top level domain in DNS: ownership of the
> right to change the definition of crucial registered values
> (the meaning of "text/html" or even "image/jpeg") might have
> some economic implications which support uncooperative behavior.
> 
> I'm not sure it's possible to address those kinds of issues
> directly, but focusing on transparency and disclosure I hope
> can help reduce some of the friction.
> 
> Larry
> --
> http://larry.masinter.net
> 
> -----Original Message-----
> From: Michelle Cotton [mailto:michelle.cotton@icann.org] 
> Sent: Thursday, February 17, 2011 12:39 PM
> To: Larry Masinter
> Cc: Alexey Melnikov; Philippe Le Hégaret; Graham Klyne; Roy Fielding; tag@w3.org
> Subject: Re: Tracking of pending media type/charset/URI registrations
> 
> Thanks for the background Larry.  I would love to explore more if these were
> IANA delays or other delays.  We are working very hard to make the process
> as smooth as possible for obtaining registrations in IANA maintained
> registries.  There are some parts that are out of our control.
> 
> I would love to work with all parties to make sure the procedures are clear
> and that the process will work for registries where there have been issues.
> If IANA ticket numbers can be provided, this will help us track down the
> requests on our end.
> 
> I look forward to talking with you all.
> 
> Thank you,
> 
> --Michelle
> 
> 
> On 2/17/11 11:43 AM, "Larry Masinter" <masinter@adobe.com> wrote:
> 
>> Hi Michelle,
>>
>> Philippe Le Hégaret in W3C was looking at complaints that
>> W3C working groups have had about not slow responses on
>> registration requests. He had some examples of places where
>> a MIME registration was in progress for years, for example,
>> and W3C was maintaining their own list of registrations
>> applied for and status.
>>
>> So at a minimum I'd look at those cases and what the nature
>> of the complaints are. Philippe, can you review these with
>> Michelle?
>>
>> In particular, the HTML working group is wrestling over
>> a proposal to use or not use IANA for "link relations" because
>> of a perception that IANA didn't "work". There's a long
>> discussion of this in HTML working group, but maybe we could
>> get some of those participants to join the conversation.
>>
>> Secondly, I am looking at some of the information that is
>> either missing or wrong or not updated in MIME registries,
>> and I have an internet draft
>>   draft-masinter-mime-web-info
>> that I'm working on that catalogs some of the problems. In
>> some cases the registry needs updating, in some cases the process
>> needs to be more transparent.
>>
>> In the case of URI scheme registration, again, there is a gap
>> in time between "name observed in use" and "registration applied
>> for" which can be years, and another gap between "registration
>> applied for" and "registration accepted" which can take years,
>> and then after the registration is accepted, no process for
>> errata or even capturing expert reviewer comments.
>>
>> Anyway, that's some background.
>>
>> Larry
>>
>>
>> -----Original Message-----
>> From: Michelle Cotton [mailto:michelle.cotton@icann.org]
>> Sent: Thursday, February 17, 2011 10:53 AM
>> To: LMM@acm.org
>> Cc: Alexey Melnikov
>> Subject: Tracking of pending media type/charset/URI registrations
>>
>> Hello Larry,
>>
>> Alexey suggested checking with you regarding the topic you brought up with
>> him described below.
>>
>> I wanted to get a feel for what you are looking for and possible examples
>> where the current process didn't work before we explore options of how to do
>> things different.
>>
>> Can you provide me with some more information?
>>
>> Thanks,
>>
>> Michelle
>>
>>
>>
>> ------ Forwarded Message
>> From: Alexey Melnikov <alexey.melnikov@isode.com>
>> Date: Sun, 6 Feb 2011 09:03:23 -0800
>> To: The IESG <iesg@ietf.org>
>> Subject: Management item: tracking of pending media type/charset/URI
>> registrations
>>
>> Dear Secretariat (BCCed),
>> Please add this management item to the February 17th telechat.
>>
>> Thanks,
>> Alexey
>>
>> -------------
>>
>> IANA (and IESG),
>>
>> Larry Masinter suggested the following:
>>
>> ---
>> I¹m wondering whether IANA might use a public tracker (or something like
>> it) to note pending registrations, reviewer comments, responses, and to
>> link the registration itself to the comments and replies. The tracker
>> could point to a mail archive if the responses were in an archived email
>> list where the archive was maintained as carefully as the registry itself.
>>
>> In general, we have situations where registrations don¹t quite meet the
>> criteria for the registry but, because the registered values are already
>> widely deployed, not putting them in the registry seems counter-productive.
>>
>> These issues apply to MIME types, charset registries and URI schemes.
>> ---
>>
>> I am wondering if this can be done easily by IANA.
>>
>> I am also wondering if this is actually modifying IANA registration
>> processes for the corresponding registries.
>>
>>
>> ------ End of Forwarded Message
>>
> 
>

Received on Monday, 21 February 2011 03:16:39 UTC