by a more general quadratic algebra (possibly obtained by deformation)
and then to derive Rq [G] by requiring it to possess the latter as a
comodule. A third principle is to focus attention on the tensor
structure of the cat- egory of (!; modules. This means of course just
defining an algebra structure on Rq[G]; but this is to be done in a
very specific manner. Concretely the category is required to be braided
and this forces (9.4.2) the existence of an "R-matrix" satisfying in
particular the quantum Yang-Baxter equation and from which the algebra
structure of Rq[G] can be written down (9.4.5). Finally there was a
search for a perfectly self-dual model for Rq[G] which would then be
isomorphic to Uq(g). Apparently this failed; but V. G. Drinfeld found
that it could be essentially made to work for the "Borel part" of Uq(g)
denoted U (b) and further found a general construction (the Drinfeld
double) q mirroring a Lie bialgebra. This gives Uq(g) up to passage to a
quotient. One of the most remarkable aspects of the above superficially
different ap- proaches is their extraordinary intercoherence. In
particular they essentially all lead for G semisimple to the same and
hence "canonical", objects Rq[G] and Uq(g), though this epithet may as
yet be premature.