RE: [xml-dev] XML-DEV list - prior art

["Len Bullard" <cbullard@hiwaay.net>]

Note that the paper that made it possible for Lockheed Martin to build
stealth aircraft was an obscure publication by a Soviet researcher.  In the
country of origin, it had been ignored.  It was recalled by an American
engineer/mathematician.   Had any of this been based on citation, stealth
would have taken a lot longer to develop.  The ideas for using flat surfaces
and fly by wire were known but the math in that paper sped things up
critically and enabled the US to take a two-decade leap over their
adversaries.

That is why one wants to be deeply aware of all of the crazy ideas out
there.  Some of them work when the right questions are asked.  

As to prior art and the patent problem, it is better to have peer review
than to depend on an overburdened small staff of government reviewers.
Otherwise, just like unfiltered search results, it is easy to game even if
expensive.

The problem of Sterling's suggestion as I understood it was cumulative flat
citation.  It could work if prior art and domain was bundled into linksets
that themselves become the reference by which claims are vetted and if that
checking is highly automated (eg, TVM indices augmented by abstraction of
logic chains that can be proofed).

len


From: Ian Graham [mailto:ian.graham@utoronto.ca] 
 
This is a fine idea, but I don't see how it could work.

The 'value' of a scientific (or any academic) paper is in the paper's 
new ideas -- and tin he references / footnotes giving it context.

Peer review determines if a paper is 'good enough' to have potential 
value, or not.

Insufficient references (or badly formed ideas) means you don't make it 
past the gate.

The whole culture of academic writing is built around this model and 
this notion of value.

But the value for code is entirely different - value is functionality, 
cost, time to market, etc ...

So unless you change the measure of value, this idea won't fly.

My 2 cents, anyway.

Ian