On Mon, Nov 19, 2012 at 3:07 PM, Mark H. Wood <[email protected]> wrote: > That ever-lengthening list of hardwired tests seems to point to deeper > problems. I'd like to suggest that the ultimate problem is that the > whole idea of Bundle is wrong. > > Containment is the wrong model. Just about the only Bundle that will > ever contain more than one Bitstream is ORIGINAL. Things like TEXT > and THUMBNAIL are really class-membership flags, and ORIGINAL can be > seen that way too. > > I think a more general solution would involve abolishing Bundle and > describing the Bitstream more directly. It could use simple flags: > when extracting flat text, for example, give it the DERIVED flag and > the TEXT flag. Or it could be done with valued attributes: this > Bitstream has DERIVATIVE("TEXT"). Or it could be a complex set of RDF > objects with relationships going off in *all* directions. This > probably needs some debate.
I agree that I never liked the concept of bundle very much, mostly because of the minor annoyance of yet another natural join when writing SQL queries. OTOH, I know people who depend on the concept with their extensions of DSpace. So it would be highly desirable to preserve backward compatibility at least on the API level. But I like the idea of items directly containing bitstreams and their relations being described in another way. The possibilities of storing RDF in a SQL database also seem to be well explored: http://infolab.stanford.edu/~melnik/rdf/db.html Regards, ~~helix84 Compulsory reading: DSpace Mailing List Etiquette https://wiki.duraspace.org/display/DSPACE/Mailing+List+Etiquette ------------------------------------------------------------------------------ Monitor your physical, virtual and cloud infrastructure from a single web console. Get in-depth insight into apps, servers, databases, vmware, SAP, cloud infrastructure, etc. Download 30-day Free Trial. Pricing starts from $795 for 25 servers or applications! http://p.sf.net/sfu/zoho_dev2dev_nov _______________________________________________ DSpace-tech mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dspace-tech
