SDD proposal:
Character and state subsets (exclusion/inclusion definitions)

TDWG working group: Structure of Descriptive Data (SDD)

Introduction

Inclusion/exclusion in general has aspects:

• Constrain item description data for reporting purposes, i. e. removing available data from output
• Constrain terminology elements available for adding new data to an item, avoiding potentially confusing characters or states

It is quite reasonable not desire unconstrained reporting output, but to constrain items available for adding new description not yet present in the item. If an excluded character or state term has already been used in an item, it is reported (it may be unexpected or even erroneous, but keeping it visible allows error to be corrected).

State subsets

In contrast to character exclusion/inclusion, state exclusion/inclusion is rarely desired for reporting purposes, whereas excluding character for reports for specific purposes makes sense. An example is the exclusion of ultrustructural, chemical, or molecular characters from descriptions aiming at field identifications.

However, preventing the display of states not expected to be applicable to a given group is frequently desired, to reduce confusion of the operator and consequently minimize the number of data entry errors and improve the speed of coding. This is especially the case in large projects where the terminology covers a diverse group with many distinct subgroups. Examples would be a terminology for:

• plants including ferns and mosses
• all insects
• all ascomycetous fungi (see LIAS)

In these cases the state terminology often contains terms specific to certain groups, which are confusing while coding other groups.

If states would be added to the presentation character grouping (in addition to characters, as it is planned at the moment) it would be possible to define editing or reporting views with selected states only. This would probably integrate seamlessly with mechanisms that have to be implemented for character exclusion/inclusion.

However, it places a significant burden on the developer, who may have to manage multiple character views. Adding a new state would require to add this state to the global definition, and enable it in all available character hierarchies or subsets.

We urgently need a discussion about how practical this is!

One way how the impact could be minimized is that builder applications provide a default operation to a add a new state to any character grouping/tree node where the character is referenced.

Note: DELTA-like include/exclude character statements are complementary during import, but have different repercussions when adding a new character. SDD currently only support the include type, where a new character by default is excluded. The DELTA "Exclude characters" directive would, in contrast, have included the new character.

Request for discussion

Please send your criticism or suggestions to the sdd mailing list or to any of the authors.

Gregor Hagedorn; Vers. 1; 10. Feb. 2003