The encoding problem (Rumelhart and McClelland 1986) is an important c
anonical problem. It has been widely used as a benchmark. Here, we hav
e analytically derived minimal-sized nets necessary and sufficient to
solve encoding problems of arbitrary size. The proofs are constructive
: we construct n-2-n encoders and show that two hidden units are also
necessary for n > 2. Moreover, the geometric approach employed is gene
ral and has much wider applications. For example, this method has also
helped us derive lower bounds on the redundancy necessary for achievi
ng complete fault tolerance (Phatak and Koren 1992a,b).