KEEPING GEOLOGIC DATABASES FRESH AND USER FRIENDLY

How are databases like a pile of dirty laundry??

Both contain bits that need to be cleaned and put away on a regular schedule or, eventually, one’s drawers become obsolete and there’s a big stink.

In the process of archiving our own and general Alaskan geologic data, the Alaska Division of Geological & Geophysical Surveys (ADGGS) has developed an enterprise Oracle database and has conducted numerous database-development projects. In our work on these projects, we have learned some important truths about databases. Possibly the most important insight is that a good database is never “complete.” Databases require ongoing care. Caretakers must continuously mine for new data, revise the structure if necessary to accommodate new types of data, perform quality control, upgrade interfaces to user expectations, and periodically migrate the databases to new platforms. Caretakers should consist of a design team well versed in data management theory and at least one specialist with skills in design aesthetics and interface usability. Databases without caretakers are at risk of becoming stale from the lack of new data or unusable due to an incomprehensible interface, poor structural design, or proprietary or uncommon software.

The single most important action necessary to facilitate database survival and growth is the creation of a clear post-development plan. Prior to committing to a database project, the agency should develop a means of absorbing the future cost of ongoing database management. Collaborative interagency databases must have a single parent agency among the group with a recognized leadership role and commitment to providing the future migration and maintenance support. Contributing organizations must secure institutional commitment to ensure that their continuing support does not wither into catch-as-catch-can style participation. The best way to ensure an agency’s commitment to a database is to integrate ongoing database administration into the organization’s normal operating budget and business process. In this regard, an enterprise database works better than a collection of smaller, fragmented, do-it-yourself, and often poorly designed databases. Clear documentation, current data, and a pleasing "look and feel" will go a long way toward encouraging staff and public interaction with the database.