Object-oriented databases (OODBs) are in many ways a better match for scientific data management than conventional record-oriented database systems. User-defined datatypes reduce the encoding going from a scientific domain to the database. Direct support for complex objects is useful for capturing hierarchical structures, such as molecules. OODBs generally have collection types, such as lists and arrays, that are a better basis than sets for the dimensional data common in scientific applications. Their inherent extensibility seem a good match for handling new kinds of metadata, and having behavior definable in the database permits transparent access to existing data in multiple formats via a common object model. We begin by recounting our experiences with using OODBs for scientific data, in the domains of computational chemistry, and materials science. The bulk of the talk, however, deals with areas that need improvement for OODBs to support scientific applications well, among them: f...

. Migrating data to a new database model presents problems if there are existing application programs that must continue to access the data, bu that cannot be converted immediately. If the target database is object-oriented, such a legacy program can be encapsulated as an object or a message. We argue that some applications will benefit from further "reification" of execution instances as database objects. We introduce a "computational proxy" mechanism and our prototype implementation of it for computational chemistry codes. We conclude with a discussion of where declarative capabilities would have been a useful adjunct to object-oriented database features. 1 Introduction Object-oriented databases provide type definition facilities that cope well with complex structures that arise in advanced applications, such as scientific computing. While transitioning datasets to an object-oriented environment can be non-trivial, providing for existing application programs to access the data in it...

Interest in database support for scientific information management has been considerably growing in the recent past. In this paper, we look into one specific scientific area, i.e., molecular biology. The analysis of the current molecular biology working environment consisting of various independent software components shows several shortcomings and thus leads to requirements for an integrated working environment. We propose an architecture for an integrated system, the centre of which is a federated database system, the data repository. Its main benefits are a central data administration, a uniform data representation, the possibility of data reuse, and ad hoc query facilities. In addition, new application tools can be implemented on top of the repository easily and thus much faster than today. 1 Introduction Due to advances in technology, instruments used in scientific laboratories get more complex and more automated. Scientists, like physicists, chemists, and biologists who use thes...