🚀 go-pugleaf

In article <1992Jul25.033617.6322@cs.su.oz.au> srctran@world.std.com (Gregory  
Aharonian) writes:
>  
>      To appreciate the magnitude of the problem of software patents, what
> follows is a list of about two hundred patents from just TWO months of 1991.
> The software community should start helping the patent office better manage
> the process of awarding software patents.

What are the criteria for generating these lists? It would seem that with some  
criteria (e.g. all methods that could be implemented using computers or  
microprocessors as one part of a larger apparatus) many more "software" patents  
are likely to have been granted in those two months.  If thes list were limited  
to methods that are performed with no other hardware than an microprocessor or  
general purpose computer, then a much smaller list would have likely been  
generated.

Mr. Aharonian's postings of neural network patents seems to be based upon  
patents when "neural networks" appear in the title or might likely appear in  
the description.  It is more probable that such patents do indeed constitute  
patents in which the application of specific neural networks technologies are  
central to the claims being generated. It is less probable that this is true of  
descriptions which contain the word "software" or "programming" in which this  
may be a peripheral aspect (such as in the process claimed in Diamond vs. Diehr   
which is mainly about producing rubber, and in which the computer plays a  
limited role). So what is PRECISELY meant by "software patents" in Mr.  
Aharonian's postings is less clear, and the need for clarity on this issue has  
been repeatedly supported by writers on all sides of this issue in this forum.
Clarification of the classification scheme used here would be helpful.

Also helpful would be some sort of tracking over time ADJUSTED by the change in  
importance, sales, investment and employment in the software industry over  
time. For example, there are many more computer users, and programmers, today  
than in earlier years. As a result, it should not be a surprise that there is  
SOME growth in patents in this industry.

It might be interesting to see the annual number of granted "software patents"   
for many years using Mr. Aharonian's selection technique.   Then it would be  
useful to normalize for the growth of the industry by dividing by the numbers  
of estimated computer users, or by the number of programmers, in those years.  
(It is probably better to count these by application date and not award date,  
as awards date might be due to factor in the PTO and not in the industry  
itself.) It would be interesting to see if this patent explosion is in fact  
growing faster, the same as, or slower than the industry itself.  Are we just  
seeing the same growth in relative importance in patents as we are seeing in  
growth of the economic importance of the software industry, or are we seeing  
something else?

Scott McGregor