Support for Abstract Data Types

Electronic notebook data include:

1. parameters used to configure experimental facilities, including templates for formatted entry;
2. operations and equipment logs;
3. result files, plots of output data, images captured from samples or processed from output data;
4. text (scientific characters), sketches, and drawings;
5. cross-experiment annotations, discussions, comparisons;
6. sequences of images, video/audio clips of teleconferencing related to an experiment.

Data such as images, graphs, and video data, cannot be modeled using the base data types supported by OPM and commercial DBMSs such as Sybase and ObjectStore. These data are usually stored as files and can be manipulated using special operations that are often available as individual software packages with their own input and output formats (e.g., various data acquisition and analysis operations).

Interacting with EN data involves using several software packages in conjunction with multiple databases and/or files. We intend to address this problem by providing a uniform interface to EN data and operations. To this end we will provide the following extensions to OPM:

1. Generic support for abstract data types, such as image, plot, and video data types--along with support for external methods associated with these data types in order to implement the data type specific operations.
2. The ability to implement an OPM schema as a collection of related databases and files, rather than a single database.

Extending the existing OPM (mapping) dictionaries will allow arbitrary new data types--along with their associated methods--to be registered with an OPM system. Such non-traditional data types could then be incorporated into an extended OPM database, consisting of a traditional database complemented by one or several files containing non-traditional data.

Furthermore, we will develop an enhanced OPM query translator for dealing with new abstract data types. This enhancement will integrate query processing with the software implementing the operations associated with the new data types. OPM queries will then be able to make use of data type specific methods, which can be implemented as external programs, possibly on remote sites.

These extended OPM tools will provide seamless, high-level access to hybrid data involving complex and highly specialized data types, and be stored in a combination of databases and other files. It will be straightforward to extend the system to deal with new abstract data types, or new software tools for manipulating or viewing specialized data types, by registering the data types and tools with the OPM system. In addition, the enhanced OPM query translator will be able to evaluate queries against hybrid databases that include calls to specialized tools for examining, comparing or viewing such data types.

The OPM extensions mentioned above will be facilitated by the CORBA compliant architecture described in section 4. CORBA supports dynamic linking of software objects, and consequently provides a general mechanism for calling external methods from within the enhanced OPM query translator.