The importance of convexity arguments in functional analysis has long
been realized, but a comprehensive theory of infinite-dimensional convex
sets has hardly existed for more than a decade. In fact, the integral
representation theorems of Choquet and Bishop -de Leeuw together with
the uniqueness theorem of Choquet inaugurated a new epoch in
infinite-dimensional convexity. Initially considered curious and tech-
nically difficult, these theorems attracted many mathematicians, and the
proofs were gradually simplified and fitted into a general theory. The
results can no longer be considered very "deep" or difficult, but they
certainly remain all the more important. Today Choquet Theory provides a
unified approach to integral representations in fields as diverse as
potential theory, probability, function algebras, operator theory, group
representations and ergodic theory. At the same time the new concepts
and results have made it possible, and relevant, to ask new questions
within the abstract theory itself. Such questions pertain to the
interplay between compact convex sets K and their associated spaces A(K)
of continuous affine functions; to the duality between faces of K and
appropriate ideals of A(K); to dominated- extension problems for
continuous affine functions on faces; and to direct convex sum
decomposition into faces, as well as to integral for- mulas generalizing
such decompositions. These problems are of geometric interest in their
own right, but they are primarily suggested by applica- tions, in
particular to operator theory and function algebras.