Grouper Developers Forum

[Tom Barton <>]

Chris Hyzer wrote:
> > > When xml-importing, it might be nice to have
> > Source.getAllSubjectsAndCache() and store Subjects in
> > > memory to avoid 'many' lookups. Retrieving Subjects using a
> > configurable size, e.g. getSubjects(Set
> > > members, int batchsize), might also be helpful.
> > I think a method in a source to get all would do the trick.
>
> Thinking more about it, I wouldn’t want all subjects in memory (well,
> I don’t not want that), but I need them in the DB so Hibernate can do
> a sorted paged query on them and not bring back everyone...  but the
> same concept, getAllSubjects, put in DB... :)  I think we are back to
> the members table having a bit more data... at least name,
> description, sortField... which is updated periodically, and which is
> not the system of record.

How an application should integrate with "identity services" remains 
wild country these days. I'm keenly interested to see how far we can 
take the Subject API approach to meeting identification needs of 
applications like grouper. I hope we'll discover that it's not always 
necessary to replicate Subject data into app-specific DBs to get 
reasonable integration.

If we know how the Subject API should be enhanced to provide better 
support for "batch" type operations, I'd very much like to enhance it.

Tom