"The most successful projects I've seen at the W3C had a small core of
talented architects, often just a single person, surrounded by a larger
group that had relatively few problems forming consensus around the
brilliant original design. In groups that lacked a cohesive architectural
center, the results tended more towards the dreaded "design by committee."

http://www.xml.com/pub/a/2006/02/01/the-power-of-no.html Micah Dubinko

Tim Bray once expressed this as the "get the right people in a room".  When 
repeating that idea, someone asked me offlist how one selects the 'right 
people'.  My answer is 'experience' informed by 'crowds of proofreaders'. 

That IS how the XML *design by committee* was set up and for the most part, 
that worked.  Feedback over a proven set of ideas is how this works.  If 
the ideas are entirely new, they probably shouldn't be on the table but 
in the lab.  An example is microformats; these are architectural forms 
come again.  Did they work last time?  Sort of.  Truth is, they weren't 
adopted widely enough to get the experience needed.  This time, the same 
wine in a brand new bottle is being distributed widely, so they are out 
of the lab.  Ontologies:  these are as old as Greek civilization and 
maybe older.  We understand this stuff.  That there are mappable 
competitors (eg, RDF and Topic Maps) means this is a validatible notion. 
That claims of recent invention die quickly in references to prior 
art means this is a safely applied idea.  That the attempts to make 
any of these work at scale aren't yet scaling means this is still a 
very competitive market and that is healthy, but keep track of what 
is an isn't working when it comes time to converge, re-invent, and 
'just say no' because...

Just Say No isn't enough.  Informed "yes" is the best approach although it
is 
usually easier to turn 'no' into 'yes' than 'yes' into 'no.

len