- From: Tim Kindberg <timothy@hpl.hp.com>
- Date: Tue, 22 Jan 2002 09:38:18 -0800
- To: Patrick Stickler <patrick.stickler@nokia.com>, ext Tim Kindberg <timothy@hpl.hp.com>, "ext Sean B. Palmer" <sean@mysterylights.com>, <sandro@w3.org>
- Cc: URI <uri@w3.org>, URN <urn-ietf@lists.netsol.com>
At 04:26 PM 1/22/2002 +0200, Patrick Stickler wrote: >I fully appreciate the benefit of guarunteed global and temporal >uniqueness (hence support for UUIDs), but I feel that there are >other means of anchoring names in time, not just by date, such >as version numbers, and there is nothing to prevent you from including >a date in an 'hrn:' root, similar to 'tag:' -- it simply isn't >manditory. But 'version numbers' don't necessarily cross organisations whereas dates do. Our view is that you have to mandate some type of framework for the coherence of the name space. > > As for the hierarchical nature of hrns -- sure, hierarchies are nice. But > > why mandate a syntax to that extent for the 'local' part of the identifier? > > It's nobody else's business; it just has to be unique. > >I'm not sure I follow you -- though per my unsure understanding, >you would like to leave everything beyond the authority portion >opaque? That's what we do. >Well, there's no restriction about having "local" syntax for >partitioning of any sort, and you can have any valid URI string >(sans '/' characters) in your local part, with just one level >of naming. But my point is: why frame or constrain where there is no global need to do so? We should agree on as little as possible (Occam's Razor for naming schemes). >Having a standardized representation for hierarchy, though, >is very useful for scoping identifier models which define the >identity of resources in terms of superordinate context. > >E.g. work/expression/realization/instance is a common hierarchy >in the library/literature fields. I'm happy to let such communities develop their own standards but it's none of my business. >I will admit, however, that the intended purpose and utility >of the 'tag:' and 'hrn:' URN schemes do overlap, and one >advantage of the 'hrn:' scheme (having of course a biased >view) is that the semantics of the 'hrn:' scheme expect an >instance of that scheme to denote a retrievable digital >resource whereas the 'tag:' URN/URT scheme may be used for >either digital or non-digital resources, which introduces >problems for software applications needing to decide what >to do with one. One of our intended use models is indeed that people can attach tags to physical entities but the point of doing so is for users to retrieve digital resources from them (http://www.hpl.hp.com/techreports/2001/HPL-2001-95.html). So I don't see the difference. Cheers, Tim. Tim Kindberg mobile systems and services lab hewlett-packard laboratories 1501 page mill road, ms 1u-17 palo alto ca 94304-1126 usa www.champignon.net/TimKindberg/ timothy@hpl.hp.com voice +1 650 857 5609 fax +1 650 857 2358
Received on Tuesday, 22 January 2002 12:38:57 UTC