Subject: Re: GIF/LZW patent
From: DV Henkel-Wallace <>
Date: Wed, 27 Sep 2006 13:31:25 -0700

On Sep 27, 2006, at 11:03 , Forrest J. Cavalier III wrote:

> DV Henkel-Wallace wrote:
>>  But the pharma examiners understand the domain better than  the  
>> software ones seem to.
> Why the difference?  That's the $64.0E11 question.  I think that
> anything that makes examiners more or less effective is going to
> exactly track the issues and reasons facing practioners.

Just what I said: some fields are older than others.

We (I?) think of the Victorians as pretty smart dudes, having built  
all those bridges, buildings (and institutions) we still use today.   
But remember that's survivorship bias.  At that time there wasn't a  
good theory of construction, or even a model of beam deformation, so  
basically people put 'em up and when one (most!) fell down they tried  
not to do it the same way.  A lot of the early theory came from  
analysis of the ones that _didn't_ fall down and kill anyone.  It's  
only now that the reverse it true.

Modern Pharma is more recent than that of course but still there's  
generally more theoretical and practices understanding _in general  
through the community_ and I think that means more among the  
examiners and attorneys too.  Its also "speciated" (subspecialised)  
so you can more easily hone in on the relevent set of experts.

Compare this to software, where hardly anyone reads the literature  
and all sorts of outlandish ideas flourish in parallel.  In fact if  
you start working seriously you fall off the literature treadmill,  
and if you try to dilligently keep up with the literature you have no  
time to actually _do_ anything.

I think this may also be happening in biotech, (but major disclaimer:  
I'm not a biotech guy.  Biotech (aka "big molecules") != pharma  
("small molecules").  I believe there may be less understanding of  
what is novel and what is not and thus a bunch of over-broad patents  
being granted over sequences and the like.  Again, this is second- 
hand and could be way off base.  But if true it supports my thesis.

So if I'm correct this will settle down in another 60 years or so.


PS: Is there a study of patent practice from 100 years ago in then- 
less-mature fields?  It would be interesting to compare the relevance  
and breadth of mechanical engineering patents than vs now: if there's  
a lot of contemporaneous overlap it would suggest poor theoretical  
understanding combined with  widespread experimentation; if not it's  
merely someone discovering something important and new.  I think the  
latter applied to the aircraft patents (which I do think held back  
development of the field).