Re: enum services registry question

Hi Harald

Thanks for the excellent summary.

However, it covers just the cases, when there is already a problem. I'd 
rather like to see _avoiding_ such problems or confusions in the first 
place. Thus, I'd prefer to make it _clear_ from the beginning, which 
specification applies. This leads to an URL pointing to a stable place of 
the spec. A stable place is certainly fullfilled in the RFC case. Also 
many other (standards) organizations have such means. But there are always 
cases, where maintaining a copy of the spec locally is more than just a 
nice-to-have. IMHO the best place for this would be IANA. To be on the 
safe side, IANA maintaining a copy of every spec related to a IANA 
codepoint looks like a good idea to me.

cheers,
  Bernie

On Tue, 8 Jan 2008, Harald Alvestrand wrote:

> Thomas Narten skrev:
>> Bernie Hoeneisen <[email protected]> writes:
>>
>>
>>> The term "permanent and readily available public specification" states the
>>> intension, but I think it is a bit vague, when it comes to what qualifies
>>> to this term. Furthermore it does not say much about the change management
>>> of such a specification.
>>>
>>
>> Correct. We can't know in advance which documents (that have already
>> been pubished somehow) meet the test, so some common sense is
>> needed. If a document is already published somewhere, and it is clear
>> that the publication is permanent, there is really no need to also
>> have a copy available elsewhere. It would be OK to do so, but it
>> should not be required.
>>
>>
>>> To be on the safe side, I guess we should ensure that a copy of the
>>> specification (as reviewed by expert) is always maintained at IANA,
>>> unless there is an RFC (that can be simply referenced).
>>>
>>
>> I'm not opposed to this, but I also think we need to remember that the
>> IETF is not the only ogranization capable of publishing documents
>> "permanently". And we should only do this if we are worried that a
>> particular document is important and might "go away" at some point in
>> the future. Is the document in question of this type?
>>
> I think there are 2 failure modes we're trying to guard against here:
>
> 1) Being totally unable to find the specification underlying a
> codepoint. In these Google-infested times, that's rarely going to be a
> problem.
>
> 2) Having 2 people holding incompatible versions of the specification
> both claiming that "THIS is the specification that was registered". In
> that case, any stable identifier (ISBN number, standards organization +
> document serial number, author + spec name + date....) will do to
> adjudicate the case.
>
> So I think the stability of the identifier is critical. Having a copy of
> the document is also nice.
>
>                   Harald