Re: [dispatch] Conceptual proposal for extensibility of Alert-Info URNs

[Paul Kyzivat <[e-mail]>]

I don't feel especially strongly one way or the other regarding the 
multiple URNs vs. single URN issue.

	Thanks,
	Paul

WORLEY, Dale R (Dale) wrote:
> From: [e-mail] [[e-mail]] On Behalf Of Paul Kyzivat [[e-mail]]
> 
>> I don't see a problem there. In the general case, if there are multiple
>> URIs then the recipient has to decide how to rationalize them into
>> something to be done. So what is wrong with giving some guidance in that
>> regard?
> 
> We could do that, but it seems to me to be opening a can of worms, in
> that the guidance would be (in the general case) how to rationalize an
> arbitrary set of URIs.  If we limit the rationalization process to the
> elements of a single URN of a particular namespace, the problem
> becomes simpler.
> 
>> Requiring all the categories to be specified in one URN has the downside
>> that all category values need to be standardized as part of a single URN
>> spec.
> 
> Strictly speaking, you don't have to define the entire set of standard
> categories in one RFC.  You do need some sort of registry for them and
> their standardized values, but you can extend the registry with new
> RFCs.  Of course, the vendor-specific extensions aren't cataloged
> anywhere.
> 
>> Conversely, consolidating multiple URNs gives the potential to
>> define a new category via an entirely new URN scheme. (I guess we
>> should discuss if that is a good thing or a bad thing.)
> 
> I'd vote for considering that to be a bad thing.
> 
> Dale
>