[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: model with overlapping variants



Edmon Chung <edmon@xxxxxxxxxx> wrote:

> > I think I like "neighbors" better.  "Similarity" sounds like a
> > real-valued function (how similar are X and Y?), while "neighbors"
> > sounds more boolean (are X and Y are neighbors?).
>
> I think both are confusing... just calling it a "variant" should be
> sufficient.

For the model I proposed I need a word that intuitively refers to a
symmetric intransitive relation.  The word "variant" sounds likely to be
asymmetric.  (If X is a variant of Y, does that mean Y is a variant of
X?  Not necessarily.  But if X is a neighbor of Y, then surely Y is a
neighbor of X.)

But it might turn out to be convenient to define the symmetric relation
in terms of an asymmetric relation.  For example, there might be tables
that tell when one thing is a variant of another (which is asymmetric),
and then we could define that X and Y are neighbors iff either is a
variant of the other.

AMC