Here's Why Software Patents Are in Peril (Fortune)

Context matters. If the judge was waxing poetic about the future of software patents on his personal blog I am not going to care if he uses layman definitions. But that is not what is going on here. He is making very technical decisions and statements about very technical subjects in a very specific problem domain. Using layman terms is not going to cut it.

Let me give you a example:

Imagine you implement a P2P virus scanning service. The way it may be used sometimes is that it executes major portions of it's logic entirely on your own computer and only starts leveraging the network of peered nodes under certain conditions, like it's using more then 20% of the available cpu cycles (or whatever).

Does that qualify for 'internet processing' or 'running on my computer processing'? Is it both or is it neither? How would the judge's judgment impact this?

What if you have a number of local machines on your local network and that local network is connected to the network and you are using your P2P anti-virus processing agents on those. The P2P protocol is smart enough to favor peers with low latency and high bandwidth locations. If it didn't qualify as 'internet processing' before, then does it now? What if the machines only have private network addresses, does that matter?

What about if the protocol is running over IPv4, but the machines have internet addressable IPv6 addresses. Does that qualify as the 'internet' now?

If I have my network of machines and absolutely no local processing of data is happening on my local computer, and I disconnect them from the internet by severing the external network connection, does that mean I no longer violating the patent because the remote machines are not part of the internet?

...

Imagine I was designing a new way of processing minerals and important part of the patent was that I was using a new type of cheap-but-strong sprag clutch with helical cut gears as part of the assembly. The helical cut is also important because it helped keep proper pressure on a thrust washer in order to reduce component cost and eliminated certain types of assembly errors, which is also part of the patent.

If I was walking down the street and shown that to people they may say that is a 'gear' or 'wheel'. If I shown them straight-cut gears they would also call that a 'gear' or a 'wheel'. If I shown them a pulley for a serpentine belt then they would also call that a 'gear' because of the indention on the surface.

So now I have to go to court and everybody just uses the generic term 'gear' during the discussion because that is the commonly accepted definition. Things would get extremely confusing very fast.

> still maintain that software patents are invalid because (1) software is intangible and can do nothing,

Patents themselves then are 'intangible and do nothing'.

Look at a patent. All a patent is a series of steps... it is a algorithm. All patents are algorithms. Software is also a way to describe algorithms. Software describes how switches in a computer operates to manipulate electronic signals going in and out of that physical computer.

When you get a patent you are not getting a patent on a particular hunk of metal or a particular machine or device anything like that. You are receiving a patent on a _process_. The patent covers the steps necessary to do something. When you see patent numbers of items it's not patented the actual _object_.. but it is a patent to do with how the object is designed, how it's made, how it gets used or something else along that lines.

I certainly do think that software patents should be abolished, but I don't think they really are unique or there is a some sort of fundamental difference that matters. All IP laws, including patents and copyrights, are completely arbitrary and are designed based on convenience and expected outcomes. There is no fundamental reality or concept that underlies this type of law, other then 'greatest good for the greatest number'.