i (mieg) strongly supports this use of the tree system by Eddy:
pondering Aberrations (again) and thought of something that
makes controlled vocabularies happen in a rather neat way. Suppose
we're really after, say,
Type UNIQUE Deficiency
Duplication
Inversion
Ring
Other
because our software needs to know which of these classes to display an
object as belonging to, but our source of information supplies a broader
(still controlled) vocabulary which, in effect, subdivides each of the
above; we could say (I'm sorry, I don't know a suitable subdivision so
I'm just using synonyms instead):
Type UNIQUE Deficiency deficient
lacking
short
Duplication multivalent
copied
polymerised
Inversion backwards
inverted
reversed
Ring cycle
circle
Other anomalous
strange
and so on. Then supplying the relevant word to an object fills in that
tag and all the intervening ones (Type and one of the UNIQUE set) so
that code which really wants one of the narrowed CV to work from can
find the right version. Users would still get the variety of synonyms
(which might, after all, be a classification within the relevant
sub-heading and, therefore, useful).
Not relevant to anything at the moment, but possibly worth sharing ...
Eddy.
