Re: We should drop the useless urn: prefix

[Peter Saint-Andre - &yet <[e-mail]>]

On 3/26/15 10:32 AM, Phillip Hallam-Baker wrote:
> Following the discussion at the plenary where I didn't get the chance
> to put the record straight, I have been looking at the existing IETF
> URN scheme for RFCs and IDs.

There is no such thing as a URN scheme. There is a 'urn' URI scheme, and 
there is a 'urn:ietf' URN namespace (and many other URN namespaces).

> I plan to add support for this to the
> output of my rfctool.

Great!

> The rfc on IETF URNs is farily old and dates from the 'wasted years'.
> https://tools.ietf.org/html/rfc2648
> urn:ietf:rfc:2648

I'm not sure what your reference to the "wasted years" is meant to 
imply. Should we throw out all work done in those years? And what are 
the dates so that we can understand which RFCs to ignore?

> So some history on URLs and URL like things. Back in 1993 I discussed
> URNs with Tim Berners Lee including the fact that to buy and sell
> 'stuff' online we would want URLs for cans of baked beans etc. So
> there should be a UPC: 'URL'.

Do feel free to write an Internet-Draft that defines and registers the 
'upc' URI scheme.

> That conversation predates the mistake of introducing the false
> distinction between URLs and URNs. From a semiotic point of view, ALL
> URIs are names except for the 'data' URI and the digest based URIs. A
> DNS name is a name. That is why is is called the Domain NAME System.
>
> Whether a URI is a name or a locator depends entirely on how it is
> used. Order baked beans from Amazon via the UPC code and it is a
> locator. You choose your 'baked beans resolution service' as Amazon,
> Peapod, Tesco, etc.
>
> Names and locators are distinct use categories but not distinct
> syntactic categories.

Well, draft-ietf-urnbis-rfc2141bis-urn aligns URNs with URI syntax, 
which should make you happy. :-)

> The distinction comes from whether the name is
> sufficiently complete to resolve the identifier to the identified or
> not.
>
>
> Since urns are not a distinct syntactic category, the justification
> for the urn: prefix disappears.

As far as I can see, your conclusion does not follow from your premise. 
Perhaps you mean that, from a semiotic point of view, we ought to have 
only three schemes (names, data, and digests)?

> It is not only useless, it is
> unnecessary. There is no circumstance in which a urn subscheme and a
> uri scheme should be allowed to have divergent meanings.

I agree that having both 'urn:ietf:rfc:2648' and 'ietf:rfc:2648' would 
be confusing. So let's not do that.

> Why make people write urn:ietf:rfc:2648 when ietf:rfc:2648 is sufficient?

Do feel free to write an Internet-Draft that defines and registers the 
'ietf' URI scheme. Although if you do, you'll violate your rule about 
divergent meanings.

> I think it comes down to ring-kissing: Lets make everyone acknowledge
> the fact that they are participating in our information universe which
> we control.

There are multiple information universes, as the existence of Handles 
and DOIs indicates. As far as I know, no one is insisting that everyone 
needs to participate in the URN universe.

However, as far as I can see, you are insisting that we destroy the URN 
universe (since it is based on a mistake). What do we tell everyone who 
has built things in that universe? Please note that the URN universe 
includes some very significant groups such as the publishing and 
information sciences community (ISBN, ISSN, NBN) and a number of our 
colleagues in other SDOs (3GPP, GSMA, ISO, MPEG, OASIS, SWIFT, etc.).

> The insistence on the urn prefix is leading to divergence where it
> comes to DOIs.

No it isn't, because the DOI folk have studiously (and, per your 
reasoning, correctly) ignored the URN universe and have built their own 
universe.

> The DOI folk understand naming at least as well as we
> do and they have no interest at all in sticking a 'URN' prefix at the
> start.

Naturally, that is their prerogative. There are no protocol police and 
the existing of URNs doesn't force anyone to use them.

> http://www.doi.org/factsheets/DOIIdentifierSpecs.html
>
>
> DOI: is a perfectly valid and well defined scheme. We should recognize
> it as such and assign a top level URI identifier.

As you are no doubt aware, we don't just "assign a top level URI 
identifier" (i.e., URI scheme) out of the blue; instead, we follow the 
process defined in RFC 4395. However, once again, do feel free to write 
an Internet-Draft that defines and registers the 'doi' URI scheme.

Peter

-- 
Peter Saint-Andre
https://andyet.com/