Re: Notes on IANA registration of variants for JMS URI scheme

Hmmm - coincidental, or useful timing:

http://datatracker.ietf.org/doc/draft-saintandre-xdash-considered-harmful/

That is, I mentioned the possibility of using variants with the prefix 
"x-", which seems to be in direct contradiction to the proposal that 
landed at the IETF in the past 48 hours.

So maybe we don't want to do that after all.

The above reference refers to the example set by an existing RFC: 
http://tools.ietf.org/html/rfc4288, which uses "vnd." prefix instead, 
and assumes registration of such types.

Given the likely objection to using "x-" in the context of the 
just-submitted I-D, I'm guessing we should be using and encouraging the 
use of "vnd." prefixes, and to have all variants registered (under 
first-come, first-served, that shouldn't be a high hurdle).

Comments?

-Eric

On 10/20/10 11:20 AM, Eric Johnson wrote:
> My action 218 [1] is to report back to the group on options for IANA 
> registration of variants in the JMS URI scheme.
>
> In case you're curious, you can look at the long list of current 
> registrations [4].
> The registration question is governed by RFC 5226 [2].
>
> One option - a "delegated" namespace - think Java package names, or 
> DNS names (where the IANA only manages the top level).
>
> Another option - one or more designated experts that will review 
> proposed registrations.
>
> The IANA defines the following "policy" definitions for ways to treat 
> registration [3]:
>
>    * private use
>    * experimental use
>    * hierarchical allocation
>    * first-come, first-served
>    * expert review
>    * specification required
>    * RFC required
>    * IETF review
>    * standards action
>    * IESG approval
>
> We can divide up the "space" of possible registered variants, for 
> example defining a prefix for variants like "x-" which defines 
> "private use" or "experimental use", and perhaps everything else comes 
> under "specification required."
>
> My suggestion:
> Since we do not expect any further registration of JMS variants, in a 
> sense, it almost doesn't matter what we choose.  It seems like a 
> combination of first-come, first-served registration, coupled with an 
> "x-" prefix for experimental use will provide appropriately low overhead.
>
> Thoughts?
>
> If we follow my suggestion, I suggest the following changes to the spec:
>
> In section 4:
> After:
> The three recognized variants (<jms-variant> above) are "jndi", 
> "queue", and "topic".
>
> ... add:
> Variant names starting with "x-" are reserved for experimental use.
>
> Also, section 9 needs changes, but depending on what we decide above, 
> that will affect what we put there.
>
> -Eric.
>
> [1] http://www.w3.org/2002/ws/soapjms/tracker/actions/218
> [2] http://tools.ietf.org/html/rfc5226
> [3] http://tools.ietf.org/html/rfc5226#section-4.1
> [4] http://www.iana.org/protocols/
>

Received on Thursday, 21 October 2010 06:20:28 UTC