Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

You can find an useful start to the discussion the following slides: http://www.slideshare.net/AndreaBollini2/dspace-standard-data-model-and-dspacecris

 

 

Meeting notes

Currently one of the main points of discussion is whether DSpace should obtain a more elaborate, and flexible, datamodel. The discussion balances on a continuum. On one hand, the datamodel currently used is very simple with a flat hierarchy. But this is also the datamodel's limitation. Certain relations can not be captured in this datamodel. So on the other side of the continuum is the datamodel's flexibility. The extend to which it can be adapted and relationships can be defined.

During this month's meeting, DSpace committer and architect of the DSPace CRIS System, Andrea bolini explained how they altered DSpace's datamodel.

DSpace's default datamodel is more or less compliant with the Portland common data model. It has a flat, fixed structure. This can not deal with issues like disambiguations.

In DSpace CRIS, the datamodel is built out of several entities which are not entirely predefined. Although there are some already existing entities, administrators can create additional entities. The datamodel also allows to define relationships between entities. This means the model is not flat, but instead hierarchical.

DSpace CRIS is also interested in following the community and keep porting DSpace CRIS to newer DSpace versions. Even Though this implies some challenges. For example, there are more and more voices coming up to develop DSpace on top of it's own REST API. This means that both DSpace and DSpace CRIS' REST API's should be aligned.

In DSpace 6 there is a large difference in authorization due to the underlying changes (use of hibernate). This will also be the case when the new Angular 2 UI is implemented. This should however not be a problem for DSpace CRIS, as it is already using hibernate.

Call Attendees

...