[YBA] i4i vs MS?

[geoffrey mendelson]

On Aug 13, 2009, at 2:49 PM, Danny Lieberman wrote:
>
> For example - IBM began listing IP and licensing royalties  in their  
> annual financial reports beginning in 2000 - about $1.5billion +/-  
> per year.  The majority of the $1.5BN is value of IP sold off by IBM  
> including IP held by divisions they sold off as well as custom- 
> development revenue.  The actual amount of revenue from their patent  
> licensing program is far less - about $125M gross the cost of IBM's  
> several hundred patent lawyers.


I just spoke to a former IBM employee who explained it to me. Since  
the time of Lincoln, the US has a law where you can project the value  
of a patent to the US government and if you dedicate it to the public,  
you can get a tax write off of half of the value.

So let's say you patent a way of improving performance of file  
systems. You dedicate the patent to the public (people of the US) with  
a projected value to the US government of 100 million dollars over the  
15 year remaining life of the patent. You end up with 50 million back  
(I assume in taxes you did not have to pay).

The question is what is the true value of that patent, both in  
prospective licensing fees, derivative works, improvement in systems  
that would not otherwise use it (BSD, Linux, Windows) and so on. Since  
there is usually inflation and you have to claim current value  
dollars, the end result looks like a big write-off, but if you had  
collected license fees it might have been much more. For example, what  
if  it would have gotten you a $1 for every computer sold over the  
next 15 years, worldwide?

It's not a myth, and it's easy to show how much they made for it, but  
it's just speculation as to how much revenue they decided to forgo by  
doing so.

Let's use a bad example. Microsoft patented the FAT file system. They  
wanted to charge a $.25 license fee for each use. If they had done so  
when other people first started to use it, they may have been able to.  
(that's another discussion). So let's assume they did. Now project  
that over the next 21- years the US government bought 1 billion  
computers that would have had to pay the royalty. So they claimed the  
cost of patent to the US government was $250M.

They could have gotten $125M in a lump sum by dedicating the patent to  
the public domain. Or they could have taken their chances and  
collected royalties on the computers.  So while Shachar calls the  
patent silly, I see it as a goldmine which was poorly managed.

Technicaly the FAT (and it's derivative's) support in Linux is a  
violation of the GPL. It was patented before it was placed in the  
code, by a third party, without a patent license or dedication of the  
patent to the public. The fact that Microsoft did not persue  
infringment does not change the reality of it.

Remember that patents do not need to be enforeced to stay valid. All  
you have to do is pay your annual fees. It is perfectly legal to let  
someone infringe upon a patent and then sue them after they have  
enough value or revenue to make it worth it.


Geoff.

-- 
geoffrey mendelson N3OWJ/4X1GM
Jerusalem Israel geoffreymendelson at gmail.com