Agenda and minutes for data model working group meeting (4/9/97)

To figure out what are the real differences between Adams's proposal and CALOO's proposal.

Minutes

Dave listed three main differences between these two models:

• Data hierarchy: flat generically typed collection of data chunks .vs. multi-level hierachy of strong typed collections.
• reconstructors and algorithms: Algorithm used to reconstruct a collection of data is an intrinsic part of the chunk, so the data can be reconstructed on demend. .vs. Reconstructors are separated from data.
• persistent pointers: the pointers between objects in different collections in the same event.

• Data hierarchy: The flat generically typed collections is more simple model than multi-level hierachy of strong typed collections. But it can only have run-time checking, not compile-time checking, which can be done for the strong typed model. Brent and Marc will produce two toy models: one is hierarchical and one is flat. We will compare both models at the next meetings.
• Reconstructors and algorithms: Most of people at the meeting are in favor of to have the reconstructors and the data separated, especially for level 3 needs. However, Dave thinks to be able to do reconstruction on demend for a chunk is useful. Dave will prepare a toy model to demonstrate his points at the next meeting. CALOO group already has their toy model ready for comparision.
• persistent pointers: It seems no big difference between these two models.