[OT] Re: Geoworks and their patent

[Lee Anne Phillips]

At Tuesday 1/25/00 09:10 AM +0000, Bill dehOra wrote:
>You might need to distinguish between an algorithm in the mathematical sense
>and the implementation of an algorithm. One the great ideas of this century
>is that you can separate the logic of a machine from its mechanical make-up.
>The entire software industry is built on this idea.
>
>Implementations of algorithms can be easily seen as machines, just because
>it doesn't have gross moving parts doesn't make it not a machine. Would you
>accept Amazon's one click patent if they used water driven gears and a
>couple of Jacquard looms?

Contrary to popular belief, the framers of our patent system were well 
aware that machines embodied ideas, which is why they allowed only specific 
embodiments of the basic ideas to be patented and not the ideas or 
algorithms behind them. So yes, I would accept a one-click patent based on 
gears and Jacquard looms, if Amazon cared to try to patent that embodiment 
of the idea. Of course they'd have to mention Babbage's prior art. I might 
even accept a specific electronic implementation on a particular computing 
machine, but there is a level of abstraction beyond which you can't go 
without strangling the source of ideas.

I'm sure we're all aware of Alan Turing and his gang of cryptographers, not 
to mention Colossus, but that doesn't change the basic conflict between the 
current state of affairs in the US and the real purpose of patent law, 
which is, as has also been pointed out by other posts, to encourage the 
public exchange of ideas by allowing someone with an idea to license their 
embodiment of it, encouraging others in turn to take that idea and either 
build on it or find a better way of doing it.

Using Turing's insight to twist the patent law into pretzels by allowing 
*any* idea to be embodied concretely, thereby allowing the patenting of the 
very ideas and algorithms specifically prohibited by the framers of 
"natural" patent law, is a result only a lawyer or accountant could love. 
In fact, if the Turing test is ever passed, you could patent people, which 
some of us might eventually recognize as reductio ad absurdum. But only in 
the USA, thank heavens.

This doesn't encourage innovation, only more and more Byzantine attempts to 
work around legal minefields laid by companies anxious to make competition 
more expensive. And the admitted incompetence of the PO makes it possible 
for ideas that have been originated and widely used by others to be 
disguised by verbiage to the point that the patent examiners don't 
recognize them, tricking them into granting a government license to bully 
small competitors with a presumption of protection and a heavy burden of 
proof and potential costs to overcome. It works only to discourage small 
innovators and protect the interests of large corporations, who can afford 
the cost of initiating a patent search and application as well as the lawsuits.

It's worth noting that MS-DOS was a cheap knock-off of CP/M, and if the 
patent office had allowed protection of computer "methods" at that time, 
Bill Gates might possibly have stayed at Harvard and been in some other 
line of work today. Whether this would be good news or bad is up to your 
own good judgement.

The fact that there are two systems of protecting innovation, patents and 
copyright, is the embodiment of the wisdom of the original framers of our 
"intellectual property" laws. If you wanted to protect a machine, you used 
patents and could protect and license the physical embodiment of the idea 
but not the idea itself. If you wanted to protect an idea, you used 
copyright to protect your specific expression of the idea but, again, 
couldn't protect the idea itself.

Ideas were always exempted from any sort of protection because the purpose 
of both laws was to encourage the free expression of and interchange of 
ideas. If you wanted to take the idea and run with it, you were free to do 
so as long as you used your own words (copyright) or mechanisms (patent) to 
do so. Copyright lasts for a long time, as is appropriate, a book can be a 
lifetime investment of research and effort, while patents are fairly short 
term, also appropriate. In fact, given the pace of mechanical innovation in 
modern times, it could be argued that patent protection periods should be 
decreased. A 17 year term made sense in horse and buggy days. I'm not at 
all sure it does now.

Copyright is an appropriate mechanism for protecting computer programs, not 
patent rights. But nobody bothers nowadays, since patents are a bigger club 
to thrash your competitors with.

Funny you should mention it, though; the doorbell I mentioned facetiously 
is, in fact, a one-click method that sets into motion a very complex 
sequence of stored program (wetware) events, determining if an occupant is 
at home or otherwise available, and eliciting a response based on private 
information known to the "doorbell" host, mediated by current information 
gathered from the "doorbell" user through past and current interactions and 
through optical, aural, or other means of recognition through a viewing or 
listening device or portal, or prearranged signal, or other information 
exchange mechanism, and selectively using accumulated knowledge to 
determine the appropriate response. Phrased more-or-less in that way, I 
have no doubt that one could have a jolly good try at patenting doorbells 
if no one at the Patent Office happened to notice the joke. Which I 
sincerely doubt they would, given their history.



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ or CD-ROM/ISBN 981-02-3594-1
Unsubscribe by posting to majordom at ic.ac.uk the message
unsubscribe xml-dev  (or)
unsubscribe xml-dev your-subscribed-email at your-subscribed-address

Please note: New list subscriptions now closed in preparation for transfer to OASIS.