Things I’m going to have to follow up as a result of DC-2007 Singapore conference:
(They potentially have implications for services currently provided by and for repositories in Australia.)
1. OAI harvesting of DC data: We need to make a distinction between creating a dublin core datastream for the oai service provider for discovery purposes and other metadata datastreams for description and other purposes. When EPrints only maps a tiny handful (4? 5?) values in a metadata record to Dublin Core oai data provider, it is smarter than the repository that tries to map as much as possible to the DC data provider — at least for purposes of OAI discovery. We need to think of the oai service provider looking for metadata records only, not for the deposited or otherwise linked articles within the repository. This sounds too obvious to mention, but keeping it clearly in mind means that we will not attempt to enter, say, dc.identifier elements that are about the deposited resources to DC datastream whose only purpose is to provide repository discovery data for the service provider. (Not that this has been a focus of the conference itself — but a result of discussions here with leading lights in DCMI and the EPrints AP.)
What is needed is metadata in the repositories on the FRBR model (which the EPrints AP has designed). — But that involves RDF relationships etc. It’s not the solution for the repository starting up this week. The other implication of this is the importance of creating extra datastreams for description et al — such as MARC, MODS et al metadata schema. (but more on MODS below)
2. MACAR work on standards for resource types is fine as far as it going, but to be really useful for the longer term future more consideration could be given to application profiles and related concepts. The current work may possibly be relevant for a setup that has too short of future. This is something other MACAR reps also here at the conference and I are currently discussing so best leave details till some of those discussions have progressed a bit further later this week. That might sound a bit worrying to those worried about deadlines, but pushing the deadlines back a wee bit could pay dividends.
3. Australia is one of the few countries that is left right out of the FAO (UNESCO) application profile that is bringing together the world’s data on agricultural resources and events — the government is already building a repository of agricultural resources and could probably do well to have it tailored to fit in with this to expose Australia’s records to the world in a way they are not currently.
4. One Australian university is about to start a repository collection of Herbarium resources. Have been introduced here at Singapore to a chief developer of science application profiles and repository metadata for agencies in this area — including a world’s major herbarium in the US. Looking forward to bridging the gaps between the local herbarium university and the resources on offer via Jane Greenberg (doctor specializing in this area and who has done much internationally recognized work already that can be drawn for this purpose) and her awarness of relevant programs and databases in this specialist area. — and ensuring that the local herbarium repository is part of an established international network if possible.
5. Understand there’s a history/archival repository planned for local history — have been introduced to major developments in cultural and archival thesauri, application profiles, that are potentially central to this when it gets under way.
6. Need to discuss with key stakeholders possibility of an Australian electronic thesis and dissertation application profile. It is long overdue — not that it will change current collections and data for Australasian Digital Theses — they can always be allocated a distinctive single or small set of properties within such a profile. And the AP Australia produces, possibly with NZ and others may also be done in collab with NDLTD and given international recognition — if so –this will potentially bring Australian theses into the same international exposure that other major international regional groupings are currently working towards.
7. MODS. Is this the beez kneez after all despite LC support? One two-part specific: The parsing of personal names into first and last names as separate elements, and the joining of affiliations with the personal name attribute — this needs to be compared with the value of current LC MARC name authorities given the little use of MADS. There are strong opinions on both sides of the fence with this. Recent and current trends re its adoption, and the positions of various stakeholders and associagted players such as DCMI need follow up. One thing is safe though: at the level at which it is used by RUBRIC partners and university repository generally it can always be easily reconverted to MARC if need be in the future. (Again, this is not a Conference topic, but has been a spinoff from conversations with leaders in the development of DCMI.)