Friday, April 9, 2010

Clarification of terminology: 'assert' vs. 'sue'

Since the discussion over all of this may continue for a while, I'd like to explain something once and for all, as as hopefully useful point of reference.

I see some commentators getting confused about what IBM was doing and how it related to what IBM had committed not to do. This is not only relevant in connection with the broken pledge (which as I pointed out in my previous post accounts for only 1% of IBM's total patent portfolio and also only approximately 1% of the patents with which IBM threatens TurboHercules). It's also important for the purpose of assessing how far IBM has already gone.

Let's be very clear on this one: "to assert patents" is a superset of "to file a lawsuit for patent infringement".

There are many who rush to IBM'S defense, mostly IBM-aligned people but also a few people whom I regard as usually very reasonable, who say that since IBM has not (yet) filed a formal lawsuit for patent infringement against TurboHercules, it has neither (i) broken its pledge nor (ii) done anything the community needs to be concerned about.

Let's start with the semantics and then look into both questions.

"To assert" is a very broad term, "to sue" is very narrow

The Free Dictionary provides a definition of the verb "to assert". In connection with rights, such as patents, it means to defend or maintain those rights. Filing an actual lawsuit is only the last resort if someone defends or maintains rights. "To assert" also includes earlier stages, such as calling someone to claim an infringement of rights. Or writing a letter to that effect. In other words, there are no formal requirements. It's a matter of common sense where an assertion begins.

By way of contrast, you cannot sue by calling someone or by writing someone a letter. To sue, you have to file a suit with a court of law. There's no way to stretch that definition.

The pledge was a commitment not to assert, not only a covenant not to sue

IBM's open source patent pledge was a commitment not to assert.

IBM wrote in that pledge: "IBM hereby commits not to assert any of the 500 U.S. patents listed below [...] against [...] Open Source [...]"

So this commitment was much broader and further-reaching than a mere covenant not to sue would have been.

No one forced IBM to say "commits not to assert". They could also have said "enters into a covenant not to sue for infringement of".

In 2005, IBM wanted to get the goodwill of the community. If IBM had only said "enters into a covenant not to sue for infringement of", people would likely have asked: You won't sue, but which other bad things do you intend to do?

IBM's letter to TurboHercules was not a lawsuit because that term is narrowly defined, but it was an assertion because that term is broadly defined and IBM has to honor its own words.

I know that some people say that IBM actually just answered a question, and that the infringement they asserted was only hypothetical. I'll deal with some of those views some other time. In this one I just wanted to highlight that IBM made a broad promise and now wants it interpreted narrowly.

I repeat myself but let's never lose sight of the fact that IBM only pledged 1% of its patent portfolio and only 1% of the patents with which IBM threatens Hercules are pledged patents. At some point we must all move on and talk much more about the other 99% of the issue (the non-pledged patents), but that one percent of the issue has a lot to do with IBM's credibility and I still see some people discussing that aspect. Both aspects are important, although in different ways.

Why waving with patents is bad enough

It's a black-and-white kind of perspective to say that as long as IBM hasn't formally sued, things are fine.

It would also be binary thinking to say that IBM doesn't want to destroy Hercules in its entirety, so things are fine.

IBM probably doesn't mind people using Hercules to acquire mainframe-related knowledge. The mainframe sectors suffers from a severe lack of trained professionals. Hercules provides a free-of-charge way to learn about mainframe technology on a cheap PC. IBM would rather control that field with its own emulators or other offerings, but this is not where Hercules really hurts them too much.

It's also possible that IBM doesn't mind software developers using emulators for development purposes. Other platform vendors are also quite liberal in that kind of context. After all, every additional piece of software written for the mainframe platform benefits IBM. I'll end this paragraph like the previous one: IBM would rather control that field with its own emulators or other offerings, but this is not where Hercules really hurts them too much.

What IBM definitely doesn't want Hercules to do is foray into the field of productive use. That's where IBM wants to force its customers to use IBM hardware - and only IBM hardware. There's an exception for the use case of disaster recovery in the z/OS licensing terms that suggests are slightly more liberal approach to that one. However, they could change those terms anytime, and what the z/OS licensing terms (obviously) don't address is whether IBM will refrain forever from using patents against disaster recovery solutions.

So what I believe IBM is doing here is using its patents to intimidate. They have sent out a signal that they draw the line where their core business interests are concerned.

They want to restrict and confine Hercules. "If you stay in the kids' corner, we'll probably ignore you. But once you leave that corner, you're in serious trouble."

In order to do so, they will do whatever they can: be it the z/OS licensing terms or be it patents.

Anyone who thinks that intimidation with patents can be accepted as long as no lawsuit gets filed should ask themselves this simple question:

Imagine you're having a disagreement with someone and that person draws a weapon and an ammunition belt with 173 bullets, would you also feel that everything's fine because this behavior is not identical to firing a shot?