The problem is in the choice of words - "win" and "lose".  It is demonstrated over and over that when presented with choice the market CONVERGES to one or two of those choices with clear inter-op between the 2.  If it does not then I argue that a mass market does not exist.  There are many businesses that actually do quite well in these markets.  One can argue that this choice is sub-optimal for many local conditions and these locales end up living with the "tyranny of the mass market".  It is not correct to say that this choice is dictated by large vendors.  The networking battles between IP, Token Ring and DecNET is a clear example where the big incumbent vendors were on the wrong side.
 
In the context of XSD vs Relax-NG, I would argue that a clear choice is yet to emerge.  XSD certainly has some momentum.  That is not to say that Relax-NG will not have traction.  The traction is unlikely to be in what we would consider mass or mainstream markets.  We can continue to rage against this choice, once the choice is made.  If the rage has some basis, someone may turn this into a lucrative business even.
 
Soumitra
 
>>Good, this is the first time we do agree.

But why this title "The Best Technologies Don't Win" and this trailing
sentence "Chalk up another win for objectivism." in your post that
started this thread, then?