- From: Butler, Mark <Mark_Butler@hplb.hpl.hp.com>
- Date: Wed, 21 May 2003 11:33:06 +0100
- To: "'Jason Kinner'" <jason_kinner@dynamicdigitalmedia.com>, www-rdf-dspace <www-rdf-dspace@w3.org>
Hi Jason, Team The document you cite http://web.mit.edu/simile/www/documents/historySystem/descriptiveNote/descri ptiveNote.pdf is the May the 1st version. There is a newer version (May the 14th) here http://web.mit.edu/simile/www/resources/history-harmony/descriptive-note.pdf Just a process note: updated documents need to be committed back to the IPSSources CVS we are using to build the website. The documents/historySystem tree is not in the CVS at all, so may disappear suddenly. See http://www.ipssources.com/ for more details. I've just requested logins for external parties (Jason and the PIs), but unfortunately this requires management authorisation so there may be a short delay. Dr Mark H. Butler Research Scientist HP Labs Bristol mark-h_butler@hp.com Internet: http://www-uk.hpl.hp.com/people/marbut/ -----Original Message----- From: Jason Kinner [mailto:jason_kinner@dynamicdigitalmedia.com] Sent: 21 May 2003 05:05 To: www-rdf-dspace Subject: Last Call for DSpace History System Descriptive Note All - I would like to propose that the DSpace History System Descriptive Note [1] move to Last Call. I would like to collect final comments during the SIMILE PI call this Friday, May 23. I will also work with Mick Bass to distribute an initial draft of the second document in the series, a design document that will describe the RDF model that will be implemented to address the issues identified in the descriptive note. This document will be made available at the PI call with the goal being to set up a review call prior to the PI call of May 30. Anyone who is interested in participating in the RDF model design discussion should contact me directly. Of the outstanding issues in the note, the most critical is that of naming and resolution. I would like to propose that we move forward in the history system with CNRI Handles as identifiers. Using CNRI Handles for all identified resources allows resolution to be approached later and to provide late binding for the contents of the resource, if they are available. As John Erickson pointed out, using the Handle system could also provide a mechanism to retrieve the metadata (in addition to any content that may also be available) for a given Handle. Although there may be advantages to using particular identifier generation schemes, such as MD5 or SHA for Bitstream objects, the Handle System provides a general-purpose identification scheme that can also be used for resolution, if needed later. A separate issue is whether the Handles used to track states or revisions should have a common base, such as that Handle for the resource. Because this technique would likely not improve search performance in an RDQL query, I'm inclined to defer this as an implementation-time decision. I'm looking forward to everyone's comments. [1] http://web.mit.edu/simile/www/documents/historySystem/descriptiveNote/descri ptiveNote.pdf Regards, Jason Kinner Dynamic Digital Media, LLC jason_kinner@dynamicdigitalmedia.com 856.296.5711 (mobile) 215.243.7377 (phone) http://www.dynamicdigitalmedia.com
Received on Wednesday, 21 May 2003 06:33:30 UTC