The hbz Linked Open Bibliographic Data Service http://lobid.org has been relaunched. The old version was built on a mix of technologies. Pubby was used for resources, a PHP script with a hard-coded SPARQL-SELECT query was used to generate the HTML-views for the organisations while the RDF-views were delivered as static documents.
Because of this mix, things were not as easy to manage as they should be, configurations where scattered all over the place, data on organisations got out of sync and content-negotiation did not work consistently.
All data is now displayed in a consistent, yet configurable way. Phresnel uses the Fresnel - Display Vocabulary for RDF in order to specify which data is selected and how it is displayed. Especially for data that includes blank nodes (such as adresses of organisations), both the human and the machine readable views are now easier to consume. Pubby's representation of these links is rather hard to understand and use, with Phresnel they are simply included.
We got rid of any language and format specific URLs. All data is now delivered via generic URLs, specific formats and languages can be requested using language- and content-negotation.
Get an English HTML view of lobid.org/organisation/DE-605
Get a turtle view of lobid.org/organisation/DE-605
In order to fascilitate the creation of RDFa-descriptions for library institutions (see here and here for background information), a Linked Data editor is part of Phresnel and thus the new lobid.org. It supports outputting various formats of RDF, amongst them RDFa. Saving data back to the triplestore is not implemented yet, so feel free to play around. Unfortunately, the language-negotation mentioned above does not work for the editor yet, so you will have to deal with a German interface for now.
Besides continuing to provide Linked-Library-Data for bibliographic resources and organisations, lobid.org will increasingly act as a prototype for Linked-Data-based cataloguing scenarios. Currently changes made in the editor are not stored yet. This will be implemented as soon as we have finalized a concept for the versioning of the data - we are looking into doing this by using git and flat files. When this workflow is in place, crowdsourcing improvements of the data is the next goal.