dbDIPview provides parallel access to multiple archived databases for non-technical end-users in the electronic reading rooms. To the archivist, it offers a way to configure and store access information. It deals with
- - representation information that supplements original data packages
- - automatic deployment of database packages (data objects) and related access packages, including redaction
- - access mechanism to the database via a browser menu with report selection and search screen for a selected database that mimics the original application.
Before export from the original environment is started, no modification of the data is needed (de-normalization, additional views, etc.). This is important if we want to maintain data integrity. As for the next step, when a database package is being ingested by the archives, an expert needs to validate it. During the process, queries are manually prepared to show the content of the tables. They are written in pure (portable) SQL and stored in an XML file. The queries also serve as proof that the information is well understood and described in the archives. The queries are stored in long-term storage as an access package for a related database AIP and thus enable future users to search the database using typical (parameterized) searches. Finally, when access is requested by an end-user, the archivist deploys the archived database and the access package. Now the user can use a menu to access the information without the assistance of an expert.
Licensing and cost
dbDIPview has been initially created in 2009 and is often being improved to cover requirements by new use cases.
Platform and interoperability
Linux, PHP, Apache2, PostgreSQL
A database can be stored in SIARD or simple CSV format. Complex combinations of packages are possible. The queries can be prepared using the information from the creator, using some dedicated query builder, or simply manually. When a basic query is ready, the search criteria screen and drill down behaviour can be added. A typical use case assumes relatively simple relational database data and excludes direct data visualization. The ordering process in the archives will result in packages, transferred to the dedicated dbDIPview server. There a database can be deployed and its access package activated. The user can be redirected via a unique code directly to the access menu for the desired database. When access is finished, the database can be immediately removed, or it remains available for other users among other active databases. The access menu lists the available reports. For each report, a search screen is configured for entering the search criteria, including drop-down fields. In the results pane, links to further reports can be available from a value in a certain column.
Documentation and user support
A demonstration example is included with a sample database and related viewer package.
End-users interact with the system through a browser user interface. Users can work in parallel on the same server on any of the accessible databases.
Basic knowledge of Linux system administration is needed for initial installation and configuration. Familiarity with basic SQL commands and understanding of database structure is needed to configure the viewer. No special skills are required for the end-user.
dbDIPview uses its XML schema for storing information about the queries, and its format for packaging the access package and optional CSV package. The queries should not use any SQL dialect to minimize compatibility issues with a target database.
Influence and take-up
The tool is in production use.
All development activity is visible on GitHub: https://github.com/dbdipview/dbdipview/wiki