Hi Mark, In a DSpace context I think a major problem is lack of agreement on, or default implementation for a storage model for various classes of content (typically aggregations). If this could be achieved then value-add services could be far more easily connected - dissemination services such as Manakin (flexible themes, theme re-use, them choice, continuous improvement of targetted themes)and greater interoperability in a broader context. The technology part is solved. The standards and the tools are (mostly) there. The consistency isn't. For example, in a DSpace context there are three journal "models" I know of: the one we developed under APSR; TDL's journal display; and FORESITE as well as probably variants on just using DC for container relations you mention below, and who knows what Fedora or other platforms do. There are probably others. I personally don't think much progress will be made in the repository value-add service space (i.e. anything other than storage and file-level access) without content and interchange models. I'm not saying there is just one (right) way to package or store journals, but if there are more than two or three I think we just get stuck in a set of unsustainable proprietary implementations which require an equal number of unsustainable proprietary solutions in order to migrate or exchange content *in a meaningful form* over time.
Fedora appear to be looking at this space in its new Content Model Architecture, but from a cursory look they are again just providing tools to achieve (a large set of unsustainable?) implementations rather than work on reference implementations for more complex object classes such as journals or other aggregations. Given the joint nature of DSpace and Fedora development and the new OAI-ORE space with its aggregations and re-use and exchange scope the technologies are available to address this but it still needs some content/metadata modelling to provide consistency for value-add service providers. Some of the work we have recently been doing involves developing METS Profiles for particular content classes which allows at least consistency in the packaging of materials for deposit to a repository (or some target process). We expect this work to continue over the coming years as we have been funded to being development of an Australian Data Commons, and this is unachievable without addressing the kind of problem you describe in you mail below. It's not a fast process. Scott. > > Message: 3 > Date: Tue, 24 Feb 2009 09:34:20 -0500 > From: "Mark H. Wood" <[email protected]> > Subject: Re: [Dspace-tech] citations, journals, volumes, issues, > articles and dublin core > To: [email protected] > Message-ID: <[email protected]> > Content-Type: text/plain; charset="us-ascii" > > May I suggest that we should never, never, never! get used to being > shocked and surprised by some of the aspects of digital libraries, but > rather to the warm feeling of having done something about them. > > It seems to me that metadata support for journal articles, while > fundamental, is the least of our worries in this area. People have > been cataloging journal articles for a long time, and it shouldn't be > that hard to map established practice to a mechanical representation. > > More urgently, I feel, we need to think a bit about how we ought to > go forward. I see two roads: > > o adapt repository software (such as DSpace) to permit the behaviors > needed in representing paper journals online > > o develop the most appropriate ways of organizing journal articles > for online presentation, and support those ways > > People are used to seeing papaer journals as *ordered* collections of > articles (buttered in-between with features and announcements and > similar matter) as issues within *ordered* collections of issues as > volumes. To go this route, we need to generalize a bit so that we can > create new subclasses of containment in addition to communities and > collections and items. Order hasn't received enough attention in the > design of DSpace. > > But is that really the most useful way of organizing journal articles? > It's important to keep the association with the journal "brand", and > it's valuable to preserve the binding to time, but I don't think that > people read volumes or journals; even if they habitually read from > cover to cover, what they read is articles and features and whatnot. > Is there really a need for volume and issue to be structural elements > at all? If we simply tag items with these attributes as metadata, and > provide top-quality tools for searching on these attributes (as well > as others), won't that serve? If we have a submission process that > can deal efficiently with repetitive metadata when submitting items > which are related, won't that be sufficient? How do readers, and > catalogers, *want* to do their work? How close can we reasonably get? > > We have a new medium to work with, and we should be careful of > accepting the needs and limitations of other media out of habit. But > we should also consider the value that we might get from old ways. > What we need is not new ways or old ways, but best ways. > > ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _______________________________________________ DSpace-tech mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dspace-tech
