Changes
a remark about the relation of Prob. 3. to Kriesell's conj. (actually a new open problem) added.
the more general problem for <math>\displaystyle k</math> terminals should be settled).
Problem 3 seems to be a weaker form of [[Kriesell's conjecture]], that would
ensure instead of <math>\displaystyle k</math> Steiner trees , a simple Network Coding solution of rate <math>\displaystyle k.</math> (Is it really weaker? I.e., can someone prove, that this would be a consequence of[[Kriesell's conjecture]]?) By [[Nash-Williams' strong orientation theorem]], a strongly connected orientation with the same property always exists. Lap Chi Lau <ref>Lap Chi Lau, ''On Approximate Min-Max Theorems for Graph Connectivity Problems'', PhD Thesis, University of Toronto(2006).
[http://www.cse.cuhk.edu.hk/~chi/papers/thesis.pdf PDF] </ref> discusses other relations of Kriesell's conjecture to Network Coding.
