[Linux-aus] FOSS patent infringements

[Paul Wayper]

Anthony Towns wrote:
> On Wed, May 16, 2007 at 10:21:41AM +0930, Janet Hawtin wrote:
>   
>> It does beggar belief that patent proponents do still promote the
>> system as a means to promote innovation and still deny that it is a
>> means to quash others in your market space.
>>     
>
> The theory is that it promotes innovation *by providing* a means to
> quash others in your market space. 
>
> In particular, it's meant to quash uncreative people that just try to copy
> what creative people do, to get all the reward for none of the effort.
>   

Indeed.  It's worth keeping in mind that there are two parts to the
bargain on patents:

Firstly, a company is encouraged to publicly declare methods and so
forth in patents because it gets protection on using them (for a time). 
The alternative is that either everything's kept secret and people
protect their ideas via skulduggery[1], or not inventing anything at all
because of the threat of a competitor copying it.  These days, the issue
is simply that the market is global, rather than local - in the 19th
century you might find your invention has got to San Francisco before
you did, but these days it'll be Peru, China or Moldavia.

The second is that the patent is now on public record and, after the
inventor has made their money out of it and its time has expired,
everyone can use it.  This is just the 'standing on the shoulders of
giants' thing.  To me, this is the key reason why patents are useful -
so that you have a public record of ideas that can be used and improved
on that builds up over time.

The only real key issue to me is that I think the current time limit on
patents - which I think is somewhere around ten or fifteen years - is
way too long.  It was appropriate for the nineteenth century, when it
could take a month to get your product distributed all across town, and
more than that to go from the test prototype to anywhere like full-scale
production.  These days, you can go from a test circuit design in a
programmed simulator to bespoke fabrication in a week, including
sourcing all the parts, and have it distributed in every country that
has a department store in a month.  With software it's even easier than
that - you can literally ship around the globe instantly.  So to me the
time limit on patents has now become their most useless feature - in a
year, you've made your money and moved on to better things.  My inexpert
opinion would be to make the patent lifespan three years, in recognition
of this.

The one criticism, of course, would be that the companies that are
making their money of licensing patents would lose out.  In which case I
would say "tough".  They're holding up progress in order to make money,
which I see as unethical.  Certainly in most cases one can argue that
their rate of filing patents has sped up rather than slowed, which would
mean that they should have new streams of revenue and therefore should
not need to rely solely on the revenue of patents older than three years.

But that's really a gedankenexperiment, with no likelihood of seeing
daylight.

[1]: a good example of this is Edison vs Westinghouse on DC vs AC at
http://en.wikipedia.org/wiki/Thomas_Edison#War_of_currents, but in these
enlightened times I'd also expect some of the more Gibsonesque
protection mechanisms (e.g. 'headhunting' being taken quite literally,
or infecting your inventors with a virus which will kill them unless
they regularly receive the antidote only you possess and isolating them
in a mesa in the Nevada desert).

[2]: i.e. companies who invent one thing, then 'improve it' by making a
minor change within the lifetime of the patent and have the patent
reissued because it's 'new'.  This is especially common in the
pharmaceutical industry, where the common trick is to come up with the
solid form and then, ten years later, 'discover' the liquid form.

Have fun,

Paul