TDWG working group: Structure of Descriptive Data (SDD)
The ability to store original data separately for each specimen is one of the priority extensions of the new standard over DELTA. Original data often have the form of repeated measurements, for which the data model used for synthetic descriptions is not adequate.
In the case of spore measurements (as well as many other measurements), two measurements are closely related. Further, a length/width ratio is calculated from the length and the width:
| Sample | Length | Width | Length/Width ratio |
|
|
|||
| 1 | 10 | 2 | 5.0 |
| 2 | 11 | 2.5 | 4.4 |
| 3 | 10.5 | 2.5 | 4.2 |
| 4 | 12 | 3 | 4.0 |
|
|
|||
| Mean | 10.88 | 2.5 | 4.4 |
Note that the average of the length/width ratio is not equal to the ratio of the length/width averages (= 4.352).
A similar case is where spore size is recorded together with septation status (1-celled, 2-celled, etc.), and the analysis requires both the total average and the separate averages for 1-celled and 2-celled spores.
In fact, the possible combinations of characters in original observation can not be foreseen, and may differ from case to case. One researcher may simply record the length and width, another adding septation, another separate lengths for each cell in multicelled (= septated) spores. A repeated observation tuple may also encompass the full set of all available characters, e. g. where several individuals are present in a collection specimen and each is described fully.
These analyses require that multiple characters can be recorded for each observation, and that these observation tuples remain connected for analysis purposes. Furthermore, it is desirable that there should be user-definable analysis functions available.
Gregor Hagedorn, Vers. 2, 19. October 2002
Earlier versions: Version 1