|
Difference between
current version
and
current version:
| At line 0 added 37 lines. |
| + !!! Use Case: Integrate data that references taxonomic names and concepts |
| + |
| + !! Actors |
| + Primary Actor : Ecologist or other scientist |
| + |
| + !! Description |
| + A scientist can integrate two data sets that contain taxonomic names or other references to taxonomic concepts, and the system resolves ambuiguities in those names, or presents the user with a mechanism to resolve the ambiguities. This basically involves processing a list of names/concepts in each of the two (or more) data sources, discovering if there is any ambiguity in the use of names, and returning a consolidated list that shows which concpet should be used for each input name/concept. "Should" can be determine by expert mappings of the concepts or by probablilistic techniques based on usage, and the user should be able to specify which approaches to use. The user can provide some context about each data source (i.e., collected in 1958 in SE Georgia) to help the system determine likely concepts in use). |
| + |
| + !! Flow of Events |
| + |
| + ! Pre-conditions |
| + * Metadata about data sources and taxa available (e.g., location, date, field guides used, etc.) |
| + |
| + |
| + ! Basic Flow |
| + # Client passes two or more sets of taxonomic names and other metadata about the taxa to the server |
| + # Server looks up usages of these names and determines if there is ambiguity between the sets |
| + # Server returns to user either: |
| + * A concept/usage that is not ambiguous for each concept passed in |
| + * A list of possible concept/usages with probability of each |
| + |
| + |
| + ! Post-conditions |
| + * Client application uses data to merge taxon information from different data sources |
| + |
| + ! Alternative Flows |
| + |
| + !! Further Details |
| + |
| + ! Non-functional Requirements |
| + |
| + ! Issues |
| + * Lots of details about just how this resolution would occur. We would like to use concept-based mappings, but as these don't exist we may develop probabilistic approaches in the beginning. |
| + * Need an API for client-server interactions |
| + |
| + !! History |
| + ; <Date> : Matt Jones, initial description of use case |
Back to Data Integration,
or to the Page History.
|
