Today we will start reading another research paper. This one is titled : DBPowder : A flexible object relational mapping framework based on a conceptual model. written by Murakami, Amagasa, Kitagawa et al. Object Relational mapping framework are well known to web application developers for the convenience they provide in programming data objects and persisting to a database. The relational schema is abstracted by an object graph that enables seamless data persistence. The authors begin by recognizing that there are many such ORM frameworks and most of them have to compromise two contradictory requirements, 1) that is the support of persistence classes that are directly mapped to relational table. and 2) the support of complicated compositions of base classes which are required by the application. These are two strong requirements and they become clearer as the application evolves. Initially a one to one mapping between each persistent class and their respective table is required and as the development proceeds, more complicated correspondences between persistent classes and relational tables are needed as the development proceeds. Foytr this reason, it is desirable for an ORM framework to be flexible enough between these two ends of the spectrum at different development stages.
In this paper, the authors do that by proposing a framework called DBPowder that addresses the difficulty in handling simple and complex correspondences. DBPowder supports direct mapping to tables with Extended Entity Relationship Model and it supports complicated compositions with ObjectView, a graph based object description form over the EER model. The EER model and the ObjctView together provide the flexibility that is required in different development stages.
Through out this paper the authors use the term simple correspondence and complex correspondence for 1) and 2) mentioned earlier. They refer to the classes in ORM as a) persistence classes that deal with persistent data in an object oriented language. b) relational schema to manage the persistent data in the RDB. and c) data conversion between object states and RDB queries and responses.
If we take the example of users and hosts, that have a many to many relationship as captured in a third join table, then we can have a simple correspondence with more complicated mappings possible by defining the mappings by hand.
In this paper, the authors do that by proposing a framework called DBPowder that addresses the difficulty in handling simple and complex correspondences. DBPowder supports direct mapping to tables with Extended Entity Relationship Model and it supports complicated compositions with ObjectView, a graph based object description form over the EER model. The EER model and the ObjctView together provide the flexibility that is required in different development stages.
Through out this paper the authors use the term simple correspondence and complex correspondence for 1) and 2) mentioned earlier. They refer to the classes in ORM as a) persistence classes that deal with persistent data in an object oriented language. b) relational schema to manage the persistent data in the RDB. and c) data conversion between object states and RDB queries and responses.
If we take the example of users and hosts, that have a many to many relationship as captured in a third join table, then we can have a simple correspondence with more complicated mappings possible by defining the mappings by hand.
No comments:
Post a Comment