- From: Patrick Stickler <patrick.stickler@nokia.com>
- Date: Thu, 9 Oct 2003 08:54:12 +0000
- To: "Eric Hellman" <eric@openly.com>, "Michael Mealling" <michael@neonym.net>, "'hardie@qualcomm.com'" <hardie@qualcomm.com>, "ext H ammond, Tony (ELSLON)" <T.Hammond@elsevier.com>, Patrick Stickler <patrick.stickler@nokia.com>
- Cc: uri@w3.org
_____________Original message ____________ Subject: RE: uri, urn and info Sender: ext Hammond, Tony (ELSLON) <T.Hammond@elsevier.com> Date: Thu, 9 Oct 2003 07:48:08 +0000 Hi Ted: > For "pure" identifiers, not > intended to trigger protocol processing (be it dereferencing or something > else), I can see the need for a small handful of schemes, based on > expectations of permanence or minting algorithms suited to different > environments. I just wanted to comment on a couple of points you raised in the above: 1. The "info" URI scheme does not make any claims about persistence (or of location independence) of resource identifiers. Any such expectations would be met by the relevant namespace authorities. One could say the same for probably any URI scheme, including http: And I would assert that the "cool URIs don't change" motto to apply as much to info: URIs as to any other form of URI's; even more so, given the nature of the kinds of schemes offered as key targets of the info: proposal. 2. Likewise, the "info" URI scheme does not provide any minting algorithms. It merely provides a means for the registration of non-URI namespaces so that identifiers minted under those legacy namespaces - whether minted by a central naming authority or delegated unto others - can be represented "on the Web". One could just as easily view the http: URI scheme, by way of its web authority component, as providing a namespace registration and minting delegation mechanism. As such the "info" URI scheme provides a bridging mechanism between resource identifiers that are "off the Web" and resource identifiers "on the Web". IMO, whether something is on the web or not has less to do with the scheme of a URI denoting it as to whether one can access representations of that thing. Merely providing a source of OWL owl:sameIndividualAs assertions equating the denotation of an info: URI with the denotation of an http: URI can make an info: denoted resource "on the web". The presumption that the info: URI has any real affect on the (ultimate) accessibility of a resource is an illusion. It only affects the ease of which resolution is possible. 3. The main functionality of the "info" URI scheme is the projection of identity onto the Web for commonly used identifiers which are well known within their target communities - library, etc. (Note that the I-D mentions that individual namespaces registered under the "info" Registry may assert a dereference capability, but we now realize that this lies outside of the remit of the "info" URI scheme and are therefore considering to exclude dereference as a potential functionality of "info" URIs.) Again, http: URIs offer all of the above, in addition to providing for (without requiring) resolution to representations and (IMO more importantly, for these sorts of term resources) descriptions. We are currently busy enabling the "info" Registry, which will allow "info" URIs to be used by many Web applications. We have already submitted an I-D (Internet-Draft) as part of the wider process to get the "info" namespace registered under the IANA Registry of (registered) URI scheme names. We remain hopeful that the draft RFC will be facilitated, considered and accepted. I'm sorry to be such a naysayer, but I see the info: URI scheme to be both unnecessary and (more regettable) a hinderance to establishing a standardised URI based means of referring to key industry terms in a way that maximises their utiity to web and SW agents by merit of providing access to their definitions directly via those URIs. I have particular concerns regarding the adoption of info: for the kinds of key terms indicated as primary targets, being involved both academically and professionally with the design and deployment of systems relating to library science, virtual libraries, and metadata driven resource management. I.e., I'm not just a URI junkie or http: groupie. The global adoption of info: rather than http: URIs for such key industry standard namespaces would have a very real-world impact on what I do. While I am encouraged that NISO and others are working to provide standardised URI based denotations for much used terms, I am very dissappointed in the form of solution being pursued. Hope that helps. It does help (me at least) understand better your support for the info: scheme. Unfortunately, I've yet to see any argument in favor of info: URIs which demonstrates a need (technical or social) that is not fully met by http: URIs. And there are substantial benefits to be gained from the (potential) accessibility provided by http: URIs which I consider a fatal omission from the info: URI scheme. Regards, Patrick -- Tony Hammond Advanced Technology Group, Elsevier 32 Jamestown Road, London NW1 7BY, UK <tel:+44-20-7424-4445> <mailto:t.hammond@elsevier.com> -----Original Message----- From: hardie@qualcomm.com [mailto:hardie@qualcomm.com] Sent: 08 October 2003 00:07 To: Michael Mealling; Eric Hellman Cc: uri@w3.org Subject: Re: uri, urn and info Hi Eric, Note that the requirements for the URN NID process are set out in RFC 3406 and that they do not require that the documentation be a standards track document. It requires review by a specific mailing list (urn-nid@apps.ietf.org) and review by the IESG. The term "IETF consensus" has been seen as ambiguous on this, but this case is very clear, as RFC 3406 sets out the steps admirably well. The IETF tree of the URI scheme registration mechanisms are set out in RFC 2717, and Larry Masinter is currently working on an update to the document to define registration procedures for other trees. There are two key issues for me in scheme registrations: can the registration adequately inform the reader where to turn for information on protocol processing based on the scheme, and can the registration adequately indicate who has change control over those procedures? Like many others, I don't see a great deal of point for the proliferation of schemes, unless the protocol processing indicated by the schemes is different. For "pure" identifiers, not intended to trigger protocol processing (be it dereferencing or something else), I can see the need for a small handful of schemes, based on expectations of permanence or minting algorithms suited to different environments. But a thousand flowers in that arena will only give us hay fever, in my opinion. regards, Ted Hardie At 6:29 PM -0400 10/07/2003, Michael Mealling wrote: >On Tue, 2003-10-07 at 17:45, Eric Hellman wrote: >> urn >> rigorous requirements but the real hurdle with urn is to get IETF >> consensus. > >Which is proving to be a fairly easy thing to do. At present we have the >following registered IDs: >IETF [RFC2648] >PIN [RFC3043] >ISSN [RFC3044] >OID [RFC3061] >NEWSML [RFC3085] >OASIS [RFC3121] >XMLORG [RFC3120] >publicid [RFC3151] >ISBN [RFC3187] >NBN [RFC3188] >WEB3D [RFC3541] >MPEG [RFC3614] >mace [RFC-hazelton-mace-urn-namespace-02.txt] >fipa [RFC3616] >swift [RFC3615] > >I submitted the 'liberty' NID proposal and the process once I submitted >it to the NID list was completely comment free. The time between request >and approval was about 1 month total. The RFC Editor will probably >publish it shortly. Its a heck of a lot faster than the MIME types >registration process. ;-) > >> IETF lapses most URN proposals and doesn't promote or use >> the ones it does. > >What do you mean by 'lapses'? All of the proposals except 'tag' and some >where the project dropped off the face of the earth have made it through >the process. The IETF is using the 'ietf' space fairly heavily, >especially as it concerns the XML registry defined in >draft-mealling-iana-xmlns-registry-05.txt. Presently the standards >waiting on is publication are simple, provreg, and sipping (those are >the ones the RFC Editor has, there are more I think). > >The identifiers have been assigned and the processes are in place. If >there is some confusion on that process let me know and I'll make sure >it gets clarified or straightened out.... > >-MM
Received on Thursday, 9 October 2003 01:54:04 UTC