Re: locally cartesian closed categories

[Eduardo Dubuc <edubuc@dm.uba.ar>]

hi,

yet another point in favor that terminal object and products  should not
be mandatory in locally cartesian closed categories:

terminal (or products) implies connection, fiber products don't.

compare with the notion of cofilter category (axiom similar to existence
of products), is connected, while pseudofiltered (axiom similar to
existence of fiber products), is not connected.

this is essentially the difference between filterness and cofilterness,
with all what it means

same thing, fiber products and not products are in the essence of the
notion of locally cartesian closedness

ps: congratulations to Bob R., I fully agree with all the good things that
were said recently about his handling of this list (not an easy job !).

eduardo dubuc