Hiding `inapplicable' character states (copy of posting to Delta-l)

Mike Dallwitz miked at ENTO.CSIRO.AU
Thu Feb 2 17:09:26 CST 1995

                                                               2 February 1995

> From: rambold at botanik.biologie.uni-muenchen.de (Gerhard Rambold)
> To: md at ento.csiro.au (Mike Dallwitz)
> You mentioned that the new CONFOR will be able to restrict the user's view
> of the characters to a subset, and merge data sets. Will it also be possible
> to restrict the user's view of the character states to a subset? E.g.
> All genera:
> #16. lower surface specials (primary thallus)/
>        1. absent/
>        2. cyphellae/
>        3. hypo-, prothallus/
>        4. cavernulae/
>        5. pseudocyphellae/
>        6. rhizines - rhizomorphs/
>        7. thallospores/
>        8. tomentum - rhizoids - spongiostratum/
> Tephromela:
> #16. lower surface specials (primary thallus)/
>     1. absent/
>     3. hypo-, prothallus/
> Miriquidica:
> #16. lower surface specials (primary thallus)/
>        1. absent/
>        3. hypo-, prothallus/
>        6. rhizines - rhizomorphs/
>        7. thallospores/
> Otherwise people would be confronted with character states that do not exist
> in the genus or even in the order (of the class Ascomycetes) under
> determination. The consequence would probably be that people then start
> thinking and trying to find character states in the material that do not
> exist at all (e.g. there are no cyphellae in the family Lecanoraceae).

This could be done fairly easily, but we have no plans to do so. A closely
related question arises during interactive identification: if some character
states do not occur in the taxa remaining at some stage of the identification,
should these states be marked (or hidden or made unavailable)? Some programs,
e.g. Pankhurst's ONLINE, do this. I am against it because it biases the user's
response towards agreement with previously entered information. If some of the
previous information was wrong, the new information is more likely to be wrong
too, which makes it more difficult to recover, whether by manually correcting
the errors or by INTKEY's error-tolerance mechanism.

Because it is generally more efficient to identify directly to species (rather
than first to family, then to genus, then to species), you may eventually want
to combine all your data into a single dataset. In that case, you will be in
exactly the situation discussed in the previous paragraph. Even if you keep
the datasets separate, the same argument applies, because the wrong dataset
may be being used because of an earlier mistake in identification to (say)
genus. Your concern that people may be confused by the `extra' states is
probably mainly for people who are unfamiliar with the group. But these are
actually the ones who would benefit most by the inclusion of those states!

Mike Dallwitz                                  Internet md at ento.csiro.au
CSIRO Division of Entomology                   Fax +61 6 246 4000
GPO Box 1700, Canberra ACT 2601, Australia     Phone +61 6 246 4075

More information about the Taxacom mailing list