The details are highly technical, and honestly I don't really understand it all either--it's been too many years since my last programming language theory course in college!
But if you're interested, check out
this page in Sun's bug database, which contains this explanation from Sun:
See
http://lampwww.epfl.ch/~odersky/ftp/local-ti.ps
particularly section 3 and the last paragraph on page 9. Admitting
type variables on both sides of subtype constraints can result in a
set of type equations with no single best solution; consequently,
type inference cannot be done using any of the existing standard
algorithms. That is why type variables have only "extends" bounds.
Wildcards, on the other hand, do not have to be inferred, so there
is no need for this constraint.
The provided link is for
"Inferred Type Instantiation Without Prototypes for GJ" (Postscript format), a paper on Generic Java (the predecessor of Java 1.5 generics). Warning: if you don't have a Computer Science background, much of the notation in that paper will probably look like gibberish to you.
I haven't had time to really make sense of the above information yet, so if someone else can translate it into a more layman-friendly description, that'd be most appreciated.
[ December 10, 2007: Message edited by: Kelvin Lim ]