[Om3] binary vs n-ary relations

[Chris Rowley]

> In the OpenMath CD relation1.ocd most of the relations are given as
> binary (e.g. eq, lt, gt, leq, geq) and in the MathML2 spec they are
> given as n-ary. What do we do?
> 
> In this case, I think that the soluton should be make them n-ary.

When I saw the order relations being n-ary my reaction was 'what does
that mean?': I may have mentioned this as a question in my earlier notes.

Although one conventionally strings infix order relations together as
if they might be n-ary I see this as a typical abuse of the notation
for relations.

To exactly what subset of S^3 does the possible 3-ary relation given
by this notation correspond?  Remember that to make general sense of
this notation we need to define this for all relations, not just
transitive ones.

Also there is no question of associativity here since, in general, (a < b) < c  
is meaningless (neither true nor false).

> I have started a discussion item on the discussion page at
> http://wiki.openmath.org/?title=cd%3Arelation1.

So I guess that is where this message sould go too.  Hmm, I tried, but
it turns out that my first trial of Swim has not been successful.
Will try and get my head around that later.


chris