I've tried to stay out of this discussion out of respect for the large amount of work the TDWG has already put into this issue, and the sentiment I've heard expressed that everyone is sick of the argument and revisiting it would be fatiguing and unpleasant. I only pipe in now because... well, forgive me.
My opinion is that while the TimBL / linked data vision may be a plausible one for the "living web" made from web sites, databases, and services that are continuously maintained and corrected, it is not so good for the archival use case, where a document is stored for ten or twenty years and then needs to be understood. Both use cases occur in the practice of science but it's important to understand that they're different and lead to different requirements. In the "living" case it does not matter so much what the identifier system (domain name, etc.) is since it can be changed if it stops working. (Persistence is not a big part of the linked data story.) In the archival case you need to be much more careful in placing your bets.
But don't let the existence of flakey http: URIs make you throw *all* of them away. http: URIs are neither more nor less viable as persistent identifiers than any other kind of identifier. The syntax of the string is not the issue; what's important is the level of commitment behind the identifier system. urn:lsid:, http: , and doi: all show extreme variability as far as that goes - in each syntactic category, some identifiers have a good chance of being understandable in the future, and some don't.
In our work at Science Commons we're using http: mainly because its resolution protocol is so widely deployed, leading to lovely unintended consequences such as the Internet Archive and Google. But persistence is hard regardless of the identifier scheme. For an example of an attempt to arrange for http: persistence for one particular application see http://sharedname.org/.
Whether the choice is urn:lsid: or http:, the important thing for persistence is to make sure some supporting infrastructure exists for the long haul. (Yes, that is a tautology.) That is not really a syntactic or technical problem.
Jonathan
On Apr 3, 2009, at 1:23 PM, Peter DeVries wrote:
I have several issues with LSIDs
- I don't agree that life science data is "different" than
geological, meteorological, or chemical data.
- The persistence of LSID identifiers is as problematic as URL
based identifiers. In fact, it is worse because the system adds an additional layer of poorly supported standards that few people understand and even fewer implement correctly.
- LSIDs increase the implementation costs significantly beyond the
costs required for domain registration and a web server. For most implementations you will need a additional machine (virtual or otherwise) and someone who understands the intricacies of LSIDs.
- Tim Berners-Lee feels that LSIDs are unnecessary, and after
spending several years looking at this issue I think he is right. Even if TBL is wrong, you have positioned yourself for an uphill battle for adoption.
- There is a well developed and widely adopted standard for
integrating data sets developed by the TBL and the linked data community that addresses the needs of the TDWG community.
For those who are not familiar with this initiative, check out the linked data site at: http://linkeddata.org/
and Tim Berners-Lee recent talk at TED http://www.ted.com/index.php/talks/tim_berners_lee_on_the_next_web.html
Respectfully,
- Pete
Pete DeVries Department of Entomology University of Wisconsin - Madison 445 Russell Laboratories 1630 Linden Drive Madison, WI 53706 GeoSpecies Knowledge Base
On Thu, Mar 19, 2009 at 5:14 PM, Garry.Jolley-Rogers@csiro.au wrote: Hi Hilmar, Struggling with exactly these issues as I implement LSID's here.
It is a concern (especially given the underlying principles of permanence embodied in LSIDS) that the LSID project itself lacks resilience. Code can still be obtained tho' the documentation I can find is out of date and dependencies may be too. I'll know very soon - by the end of today. Like many things out there.. It seems to be withering now that the initial enthusiasm has died. While the collections community may think in centuries, permanence in LSIDS seems to mean a few years. Perhaps my google-fu has failed me.. If so please tell me.
My questions.... Is there sufficient interest and community involvement to keep it alive .... Even it is no more than an update documentation & co. Perhaps it should be brought into the TDWG fold? Any comments? Happy to contribute what I can.
GarryJR
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Garry.Jolley-Rogers@csiro.au Biodiversity Informatics, Taxonomy Research & Information Network Centre for Plant Biodiversity Research, CSIRO Plant Industry, GPO Box 1600, Canberra ACT 2601 AUSTRALIA w:(02) 62465501 http://www.cpbr.gov.au/cpbr/staff/jolley-rogers-staff.html .·'¯`·.¸ ><((((o> .·'¯`·.¸¸.·'¯`·.¸ .·'¯`·.¸¸.·'¯`·.¸ >=}}}}}}/o> ><((((o> ><((((o> -----Original Message----- From: tdwg-tag-bounces@lists.tdwg.org [mailto:tdwg-tag-bounces@lists.tdwg.org ] On Behalf Of Hilmar Lapp Sent: Friday, 20 March 2009 5:07 AM To: Technical Architecture Group mailing list Subject: [tdwg-tag] SourceForge LSID project websites broken The websites for the two LSID projects on SourceForge are broken: http://lsids.sourceforge.net/ http://lsid.sourceforge.net/ I believe the latter project is defunct (can someone confirm this?) but the first should be alive, right (and this URL is in fact linked to on the TDWG website). Does anyone know what's going on? -hilmar -- =========================================================== : Hilmar Lapp -:- Durham, NC -:- hlapp at duke dot edu : =========================================================== _______________________________________________ tdwg-tag mailing list tdwg-tag@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-tag _______________________________________________ tdwg-tag mailing list tdwg-tag@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-tag _______________________________________________ tdwg-tag mailing list tdwg-tag@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-tag