Thursday, May 03, 2007

Astonishing announcement: RDA goes 2.0

As a result of a meeting at the British Library on April 30 and May 1, it has been announced that RDA and Dublin Core communities will work together in the following ways:

  • develop a formal RDA Element vocabulary (probably following the DC Abstract Model)
  • develop an application profile of RDA for Dublin Core using FRBR and FRAD
  • use RDF and SKOS to disclose RDA vocabularies (see Diane Hillmann's work on this)
The document lists the "benefits" as:
  • the library community gets a metadata standard that is compatible with the Web Architecture and that is fully interoperable with other Semantic Web initiatives
  • the DCMI community gets a libraries application profile firmly based on the DCAM and FRBR (which will be a high profile exemplar for others to follow)
  • the Semantic Web community get a significant pool of well thought-out metadata terms to re-use
  • there is wider uptake of RDA
So it appears that the call for a modernization of the library approach to metadata has been heard. What does all of this mean? Diane Hillmann was at the meeting, and I asked her the following questions:

KC: What does it mean that there will be a "formal RDA Element vocabulary?"

D: It will look something like the Dublin Core registered terms. They will be both human readable (as displayed in a browser) and machine readable (in a format like RDF). Try clicking on this link, and you can see on the right the different machine-readable formats.

KC: What happens now to the "tome" that has been developed through the RDA process?

D: The "instructions" as we see them in the RDA documentation, will not be affected. The element vocabulary, the formal vocabulary, will be separated out, and the documentation will point to the formal vocabulary terms. Many users of the documentation will not see the formal element vocabulary and may not know that it exists. The vocabulary, however, will be behind the online tools that are being created. This will make it easier to create a system that allows people to click on a term and get a definition or to see the related hierarchy.

Having the formal vocabulary means that there can be a testbed for the many and complex relationships that are being expressed in RDA, FRBR and FRAD.

KC: How is this going to be accomplished? It looks like a lot of work.

D: There will be an effort to find funding to accomplish this, but the work itself will be done with the combined efforts of the RDA and Dublin Core communities. Some of the work, such as the RDA vocabularies, is already begun.

This is nothing short of revolutionary, at least in comparison to where we started in the late 1990's on a revision of AACR2. Imagine a library that is seamlessly integrated with the semantic web.... we seem to be on our way.

12 comments:

Mark J. Andrews said...

Well, either Hell froze over or there is a God. Or both.

This is good news, especially the hook into the SemanticWeb and RDF.

Simon Spero said...

If you see a JSC member, hug them.

Irvin Flack said...

I think this is good (and big!) news for both DC and RDA, and will get librarians actively involved in DC again. I can see the DC eprints application profile work has paved the way for this extra step.

One q: 'The "instructions" as we see them in the RDA documentation, will not be effected.' Shouldn't that be 'affected'? If not, the cataloguing community will be _really_ up in arms!

Karen Coyle said...

Irvin, yes, thanks, you are right, I effected the changed to affected :-). I think this is a grand step forward and it should make library metadata much more visible on the web and also provide a boost for DC.

Bruce said...

Karen: really good to see this, though as I mentioned on another blog, it would be a really good idea to get Ian Davis (from Talis) involved in this work. He has the kind of pragmatic RDF (and FRBR-in-RDF) background that can be really helpful in this sort of work.

Alison said...

For those of us who currently do "normal" MARC cataloguing for traditional OPACs: what does this mean? Does it just impact electronic resources (e.g. ability to embed standard metadata that libraries can easily use) or does it impact exchange of bibliographic information for print, A/V etc. materials also? I'm not a techie... :-)

Karen Coyle said...

Alison,
Since this is something that has just happened, there are a lot of unknowns. However, here are a few "knowns":
1. this is a change to the creation of the standard; it means that there will be a machine-manipulable version of RDA. It does not directly affect the creation of catalog records, although eventually those could also change.
2. it will take a while for this change to happen, and even longer for it to have a ripple effect on library systems. My guess is that it will be years (but hopefully not decades) before the library system itself is using data in a different format.
3. The end user, and the cataloger, may never see any major change in how the records look on the screen. This is "deep background" stuff. However, in the best of worlds, the end user and the cataloger views will improve because library data will be more flexible.

Steve Toub said...

I'm pretty ignorant about RDA and MARC and am still having a hard time seeing the big picture here.

Can you connect the dots for me: how will the ability to express RDA elements in a machine-readable way affect and interrelate with data structure standards like, say MODS or MARCXML?

Karen Coyle said...

Steve,

Believe me, these aren't easy concepts, so I think Diane and I need to work on a follow-up document that spells out the impact of this news (assuming we can put it into words ourselves). Essentially, it doesn't directly affect MODS or MARC or MARCXML as long as we keep doing those in the same way we always have. But if we move forward toward MARC2.0 and MODS2.0, then we are much more likely to be working in an environment where 1) there is greater compatibility between metadata schemes 2) there is more re-use of the data elements that libraries define (say for personal bibliographic software, or citation linking). It essentially makes it more possible for us to play in the greater Web sandbox. The flip side is that it holds out some promise to end the isolation of library data from the "bright web." But we'll have to make some changes for that to happen.

Neil Godfrey said...

What might this mean for those seeking to establish metadata standards (including element names, thesauri for resource types . . .) for repositories in Australia?

Karen Coyle said...

Neil,
I think this is significant as an example of how to make our various metadata efforts more compatible. If we can reconcile DC and RDA, we should be able to also begin to bring into the fold other metadata standards. The DC-RDA work will be a good test of this concept. In particular, I think that the concept of a registry of data elements is key, and will allow us to re-use data elements across a wide range of metadata schemes.

Brigid Nischala said...

This is certainly welcome news. I was a bit anxious about RDA, as the early drafts didn't seem to have anything more than semantical changes. I'll be watching this development closely. Thanks for posting it!