2.0 Unit Metadata

The typical use cases for Unit metadata include:

1. Search for a unit given a parameter e.g. name, description, input/output types etc. e.g. a user might know what they want to do but they are not sure of the name of the Triana unit that can perform this type of operation. If we associated a semantic description with each unit then this type of search can be realized.
2. To point out incompatible types when connecting units;
3. To enable automatic translation from a task graph created in another dataflow or workflow environment into one that would be equivalent within Triana.
4. To allow the user to refer to a tool by a pseudo name, not necessarily the class name.
5. To enable the users to use the most recent version of the unit. Versioning within units is important as it enables programmers to get the latest unit code and users need to know which version a unit is so that their results can be reconstructed accurately. There is no guarantee that newer versions of units are backwardly compatible (we would hope that they are though) so along with the main data flows, we need to reconstruct the algorithm of units along with their correct versions.

The use cases motivate a number of requirements for unit metadata:

1. Name of unit and id;
2. Number of input/output nodes;
3. Description;
4. Help files name and location;
5. input/output types;
6. Parameter name/value;
7. Class name;
8. Version no.

Below is an XML example of this information: