conjectured in 1968 that the edges of a connected graph on n vertices
can be decomposed into at most (n+1)/2 edge-disjoint paths. This
is still open, even for planar graphs. In this talk we will discuss some
related results and special cases where it is known to hold.