Steve Mills, senior vice president and group executive for IBM"s software division, discussed open-source licensing issues and patent reform during an interview with Computerworld here at the LinuxWorld Conference & Expo. Part 1 of the interview -- in which he discussed Linux usage trends, IBM"s involvement in the development of Linux and open-source virtualization software, and the company"s plans for AIX.
Part 2 of the interview:
One of the topics that cropped up during a couple of keynote speeches here is the need to stop the proliferation of open-source licenses. Stuart Cohen [CEO of Open Source Development Labs (OSDL)] said the Open Source Initiative recognizes 57 open-source licenses. What"s your take on that? I haven"t counted them, but if there are 57, I"m sure there"s a lot of duplication inside of that 57. That incorporates a lot of subtlety. I suspect that inside of that, they reflect a few different models of licensing. There"s always going to be more than one. I mean, it"s hard to imagine that the world"s going to get to one licensing model. So we"re not necessarily bothered by the fact that there are multiples. We kind of expect that.
What about the call for the GPL? I guess this was HP"s thing [on Tuesday]: Everybody should adopt GPL. Well that"s never going to happen. There are multiple viable popular sets of terms out there. ... You can net this down to two major camps. One is a GPL license, which is very prescriptive in terms of what it means to package things with GPL and the obligation to deliver open-source for anything that is literally packaged with a GPL-licensed product.
We use the Apache license as an example, [and] there are many derivatives of that license. That"s actually the more common licensing type for open-source in the industry. Our license looks very similar, as do other licenses. That license does not carry the same restriction on delivering everything. So if you have code that incorporates that license into a product, you"re not obligated to deliver everything else in the product as open-source.
Didn"t the IBM-created Eclipse project have its own licensing model? It"s an Apache-like model. It"s not a GPL-like model. ... You get different dot-orgs, you"re going to get different derivatives. It"s hard to say to any independent body that they can"t have their own licensing. ... Look, every software vendor in the industry has a unique license for their products. Customers are used to dealing with these things every day. The fact [is] that there are a bunch of open-source licenses on top of tens of thousands, 60,000 other licenses that come from the commercial software companies. ... Every person you deal with in this world will tell you that [there are] too many contracts, too many lawyers, too many -- it"s the world we live in.
Do you disagree with what Sun did with open-source Solaris? They"re hunting for ways to grow their business. I assume that"s why you do those kinds of things. I"m not sure where that particular strategy goes in terms of what that"ll mean to their financials as a company. But the fact that there are multiple licenses out there, to me, this is sort of a silly issue. Oh, why are there 900 models of cars? Or, why do they come in different colors? I mean, it"s just nuts. Human beings are pretty creative. We like choice. We come up with these ideas. We see things in different ways. And if they make business sense, economic sense, if they deliver value, why not? That"s variety.
So you break from HP and the OSDL, which both complained about the proliferation of open-source licenses issue during panel discussions. I think it"s wishful thinking. Look, you could always say, "Gee, if these things are similar, why have as many?" If there are 57, and 40 out of the 57 look very similar, you could sit around and say, "Look, why can"t we just turn those 40 into the same license?" If all 40 entities that own those particular licenses want to get together and converge them, because they are essentially the same, hey, great.
But you"re not going to work on that. No. I"m not going to work on that. It"s not worth my time and energy.
What are your thoughts on the need for patent reform? We have been very much involved in patent reform. Our view in the world of software is that we are seeing a proliferation of patents that we would characterize as nothing more than the patenting of standard method and procedure. The overwhelmingly dominant use of software is to codify a business process and to make it therefore repeatable within the computer. Instead of having people handle something, let"s have the computer do it. There"s a process. There"s a procedure by which it"s done. Look at any of the classic paradigms of business in terms of the way purchase orders are structured, the way in which ledger posting is done. You can go on and on and on. We essentially look at the real-world physical thing, and we say, "How do we digitize that? How do we create a computer analogy of a real-world process or procedure so we can make it consistent and repeatable?" That"s what software has been doing for as long as there"s been software.
You look at those things and you say, "Should that be patentable?" It"s a derivative of simply the filling out of a seven-part form that we used to do back in the 1950s that we"ve now automated into a computing procedure, and somebody has gotten a patent on that. It raises a lot of questions about, Is that really a valid invention and how is that patent then going to be asserted by the company that owns that patent?
We have instances of companies being in business solely to own patents and assert patent rights, having no commercial product interest. Companies that hold intellectual property [typically] agree to cross-license each other for purposes of both companies having freedom of action to operate in the marketplace. There"s no basis for an agreement with somebody who just owns a patent on something.
But there are a lot of valid and important reasons to have patents in the world. There are real inventions and real creations and things that we see value for based upon their invention. What"s happening in the world of software is that the line between real invention and nothing more than somebody taking textbooks and turning them into patent disclosures has now muddied up the world of software patents and made this a very complicated and difficult situation.
We"d like to see patent reform in the form of better policing of interface patents, simple method patents, things that are questionable in terms of whether the patent offices around the world should actually be granting patents. And certainly the opportunity for a patent to be publicly posted and vetted in advance of being granted would be a nice change in process. [It"s] something that has been discussed in Europe, a fair amount in Japan. It"s getting some discussion in the U.S., but it"s not part of today"s process.