More patents

by Miguel de Icaza

Seth has raised some valid points on the patent problem, as it related to Mono.

A big problem with everyone raising potential patent problems with Mono is not that they are wrong in any way, but there are a few problems with it:

  • Not using Mono in any shape or form is not a blank waiver against patents. That means that even if you choose to stick to your beloved C, Python, C++ or anything else, for any new software you write, you are likely to infringe on someone else's patents (or even the same ones that Mono could potentially infringe).
  • Patents can be: declared void with prior art showing that such invention did exist in the past. Alternatively, you can route around it by using a different technique than the one described in the patent to provide the same feature (or something close) to avoid infringing on a patent.
  • Most people have no idea of what a patent is, or how it works. The actual invention is the claims bit. Not the introduction, not the summary, not the background, not the references and not the drawings. They help in making the claims, but they are not the patent itself. If you want to play patent-lawyer on Internet, you should at least familiarize yourself with the process.

Now, it is hard to argue with the nay-sayers about routing around the patent for two reasons: we do not know what might be patented and valid (ie, no prior art can be found or properly articulated) and most importantly for any given topic we can engage in weeks of discussion on what-if scenarios.

We do not plan on infringing patents, and if were to infringe on patent, we will either find prior art that renders that particular claim invalid, or rewrite the code to work in a different form. We do not like software patents, but we will abide by the legal rules.

If you can not think of a different way of running a C# program than the one that exists today, you are not a very imaginative/innovative programmer. The worst possible scenario is not `They will stop distribution forever'. The worst possible scenario is `They can stop distribution until we find a workaround'.

And again, remember, the software patents problem is not limited to the specific instance of Mono. Everything Seth said applies equally well to every bit of our open source stack today: do we infringe on a Microsoft patent? Do *you* know for sure you do not? Have you performed a patent search? On every possible bit?

Red Hat has chosen to adopt Java (despite the same potential problems with patents) and has decided that it is in their interest not to use C#/Mono. Like Red Hat's Seth states: this is self inflicted damage on their part, and they will not be able to ship any of our leading edge GPL code (Simias, iFolder, F-Spot or any of our future development tools).

Red Hat could either stop whining, and have their developers work in Mono and use Mono, and help us fight bogus patents or route around them, or they can keep posting to their blogs more fear-mongering.

Andy Satori has posted his insight here.

If you are going to reply, I just ask you that you take a step back, and for every instance of the word `Mono' replace it with every major open source project today `gnu libc', `linux kernel', `Open Office', `samba', `x11', `cairo', `gtk+', `qt', `binutils', `gcc', gnome', `qt', `mozilla', `my favorite file system', use your imagination.

Does your foe have a patent to it? Or someone that can be acquired by your foe?

On Stop Energy: my policy

Most people operate in Stop Energy mode, so I typically ignore them, and keep moving.

A small story I like to tell people: when I started writing Gnumeric, I was very afraid of one thing: the computational engine. How do we recalculate the value of cells when a change happens? How do we make this perform well? How do we do iterative computations? How do you resolve recursive references?

All of those problems were fairly scary, and I did not have an answer to them. I looked at all the source code I could find for spreadsheets around that time, and none of it did even a remotely good job: it was all pretty amateur, and none of it really did anything remotely close to what commercial software did.

I started work on Gnumeric nonetheless, figuring `When the time comes, I will face that problem', and spent the next three months making sure that Gnumeric was visually pleasant, that it looked like Excel, and that the "feel" was right. I tried to implement computations trivially during that time in a couple hour hack and that failed miserably.

By the third month, I decided I would not touch a computer until I figured out an algorithm for doing these computations, I took a pencil and a notebook and went to write down the steps. Surprisingly after a few hours of work I had something that looked correct.

That same day I implemented the computational engine with the features I wanted and it just worked!

What I like about this story, is that I could have given up at any point since there was a large problem ahead of me: a problem I had no answers to. And I see this with many free software developers, students and even in normal social situations: people stop doing things because they see a big problem ahead of them that they can not possibly conceive working around. My advise to every young programmer is to start writing code and delay addressing imaginary problems until they become real.

I like people who find and propose solutions.

The Mono team (both the community and the Novell employees working on it) is pretty much such a group: a group solving problems, and moving forward.

Interview with a soldier

Ran into this interview this morning. From the Sacramento Bee.

Incidentally, watched For of War this week, McNamara at some point says `We need to win the hearts and minds of the Vietnamese people'.

Posted on 20 May 2004