Re: Inevitability of ordering products

Eduardo Dubuc wrote:

what sense has the concept of unlabeled graph ?

try to put an unlabeled graph inside a computer ?

you mean unordered? i would implement it as an ordered graph, with an additional involutive map on the edges, ie Edges <--inv-- Edges ==dom,cod==> Nodes dom.inv = cod inv.inv = id --- which, in a way, confirms that

well, unlabeled graph has to be a quotient by an equivalent relation ...

isn't the "ordering" of the components of a product AxB (by the names, colors A and B), in a similar way, "factored out" by the canonical isomorphism with BxA? isn't coherence theory the way we can always factor out such arbitrary annotations on objects? (SORRY i am posting too much.) -- dusko