Archiving of Scientific Data
Many scientific databases have the properties that they are accretive
(new information is added, but relatively little is deleted) and that
updates to existing data are relatively infrequent. Keeping archives
of such databases is essential for accuracy in scientific citation.
We present an archiving system with the following properties
The system relies on the fact that scientific databases are typically
hierarchical and have a hierarchical key structure. It can also use
"diff" files when no such key is available.
- It can hold many versions in a file that is little larger the most
recent version of the database
- A specific version can be extracted by a linear scan of the this
- Queries, edits and updates can be performed directly on the archive
- The system works well in tandem with XML compression to achieve
- It provides a natural representation of "object histories"
Peter Buneman Sanjeev Khanna Keishi Tajima Wang-Chiew Tan