I spent Monday in London at a meeting of the working group of a JISC-funded project which is developing a Dublin Core application profile for the description of images. The project is led by Polly Christie and Mick Eadie, both from the Visual Arts Data Service, University College for the Creative Arts, and is one of a number of "sibling" projects funded by JISC to develop DC applicatiion profiles for various resource types.
Despite the best efforts of First Great Western to the contrary, I got there in time to give a short presentation on the concept of the DC application profile from the perspective of the DCMI Abstract Model, summarising the current work on the Description Set Profile model and introducing the "Singapore Framework".
My presentation was really just background for a presentation by Julie Allinson from the University of York, who described the experience of developing the ePrints/Scholarly Works DC Application Profile.
It turned out to be quite an enjoyable meeting - it felt to me like everyone in the group "engaged" quite enthusiastically with the discussions, which is always a good sign (and is spectacularly impressive for a Monday morning!)
I haven't really worked much with metadata for images, and I'm not that familiar with the models in use in that domain. Polly and Mick circulated a draft model based on the VRA Core, which made a primary distinction between the types/classes Work and Image. This prompted a good deal of discussion, both from the viewpoint of whether that model really addressed all the use cases at hand (e.g. Does it handle the "born-digital" case? And if a digital image in a scientific publication is generated from data, what is the (VRA Core) Work?), and also as to how well it "fitted with"/mapped to the FRBR model (on which the ePrints/SWAP profile was based)- the VRA Core concept of "Work" is not the same as the FRBR concept of the same name. This in turn raised the broader question of whether these various DC application profiles should be framed within some shared, over-arching model.
On re-reading the introduction to FRBR this morning, I note that the section on "Scope" does state:
The study endeavours to be comprehensive in terms of the variety of materials that are covered. The data included in the study pertain to textual, music, cartographic, audio-visual, graphic and three-dimensional materials
so at least some classes of image were considered as in scope by the developers of FRBR. I'd be interested to receive any pointers to/comments on any experiences of applying the FRBR model to graphical resources. I'll forward any comments received here to the project.
As I say, an interesting and enjoyable day - though I suspect for Polly and Mick it may have been one of those occasions where they left the meeting with the feeling that the task facing them was even broader and more complex than they had envisaged at the start of the day!