Re: [iri] 4395bis #54 (new): Give more details about what 'provisional' status means

On 2/9/2011 9:23 AM, Mykyta Yevstifeyev wrote:
> Hello all,
>
> Since I have initiated this issue, I'd like to stateit more detailed 
> and propose some ways to resolve it.
>
> Firstly, what is the problem?  RFC 4395 mentions that there is a 
> separate category 'Provisional' URI schemes.  However there are no 
> explicit definition of its purpose.  Therefore, the community members 
> have great misunderstandings with this status and this makes some sort 
> of problem.
>
> So I'd like to propose the following solution.  Firstly, set the clear 
> criteria for Provisional URI schemes - that URI schemes are placed in 
> the Provisional category to indicate the intention to define it more 
> carefully later and perhaps set the time it would be deleted from this 
> category if no acceptable specification appeared (maybe 6 to 9 
> months).  This novelty will allow to revise the Provisional 
> registrations if no specification of it are available.

I can think of a number of reasons that an "acceptable spec" never winds 
up appearing, yet the scheme continues to be used. Such notices should 
not disappear "automagically".

Sometimes provisional is a means to give people notice that a scheme is 
being used "in the wild" -- there *is* no public and/or acceptable 
specification, but there is a sufficient user population that someone 
trying to create an "official" spec using the same name may run into 
difficulties. Such notices should not disappear "automagically".

Sometimes schemes are defined by proprietary entities, and there really 
is a full specification for the scheme, *but* it's not publicly 
available, so it can never be registered in the Permanent registry. But 
it can be put into the Provisional registry. Such notices should not 
disappear "automagically".

I think of the Provisional registry as more of a "caveat emptor" 
registry -- the schema listed there are not well defined and may never 
be well defined, so be wary of the names listed there.

> Moreover, I'd like to ask to make a notice in your document regarding 
> 'afs' URI scheme.  I'd like to propose to mention its classification 
> as Historic directly in the document to resolve all the issues 
> concerned to 'reserved' RFC 1738 schemes.

I would leave that to the IRI working group to decide by consensus.

     Tony Hansen
     tony@att.com