Coding characters

[Mr Fortuner connection modem]

Is writing pink rather than #6,1-3/ a matter of personal taste?

Actually,
there is more to it than taste.

The Delta format, and others like it, have
been developed by people used to dichotomous keys and it was designed mostly
with dichotomy in mind. This gave it a particular slant which makes it well
suited for dichotomy, but not as well suited for anything else.

Dichotomy,
even in the modern guise of multiple entry keys, typically is a sequential
process, i.e., the characters are considered one at a time, each independantly
from the other characters. There is no need to make any general statement
about the characters.

The problems show up when we want to move from
dichotomy to other types of identification methods, or worse, to use the same
data for other purposes in other disciplines.

How would you enter metadata
for exemple, even if new fields were added to the Delta format? For exemple,
how would you say that all color characters are ambiguous?

Let's use the
flower exemple again:
  #6. flowers/
           1. white/
           2. blue/

          3. red/

First, notice that the concept of "color" has never been
specified when the character #6 was defined. Humans would understand but the
computer would have to be told.

Second, you cannot tell the computer that all
characters #6 are ambiguous, because in another data set, #6 might mean "leaf
length", and length is not an ambiguous character. You would have to enter the
metadata individually for each character and for each data set. It can be
done, but it sure is klunky.

The Delta standard was great when it was
created, 15 years ago. Computer science has changed since that time. A lot.
Now, we have object-orientation, artificial intelligence, expert workstations,
and we need a standard which will answer the needs of the latest
developments.

Renaud Fortuner
fortuner at math.u-bordeaux.fr