Thinking about software

I’ve spent a large part of the past 20 years working on university-originated software IP–in research, instruction, and administrative services. We’ve done open source, source available, venture-backed start ups, technology access programs, commons, publication agreements, distribution agreements, site licenses, waived rights, built projects with five and ten and fifteen year runs, generated $30m or so, helped give away software, release data, and build goodwill.

Often seen by university technology licensing offices as a subspecialty of a subcategory of technology transfer, software actually represents a tremendous opportunity to understand technology transfer in a different light. Rather than looking at software management as a narrow part of the application of patent rights to “physical sciences / engineering” inventions, software forces a generalization of technology transfer away from patent licensing (which still happens) to the broader question of how to augment the impact of information assets.

Before an invention is the subject of a patent application, it is an information asset. Without the patent application, the value of an invention would be its information value. The information value doesn’t simply vanish when there’s also a patent, unless the practice around patents fails to realize that value. That is, too particular a focus on licensing patent rights might as a consequence of trying to do one’s job according to policy, result in diminishing the information value of research inventions.

In this, one might see that the actual first step in university technology transfer often is to try to transfer, often implicitly, the information value of an invention into the package of a patent right. One sees this transformation expressed, later, in a patent license in which the patent is given the value, and the interaction with the lab and its inventors (and others) is thrown in, perhaps as “know-how” but often simply as an assumption that it is in the inventors’ interest to help since they stand to receive a personal share of royalties (at some often distant point in the future). This, even where the information value early on may be at a premium, well beyond the value that would be established with sponsored research cost accounting.

The business question for a new research information asset is, for the university, who to teach the information to, and how. The more specific question of technology transfer then is: “is what we distinctively know from our research efforts worth teaching to others?” Not nearly so much, “does this have commercial value realizable through patent licensing?”

It’s not that commercial value isn’t important. It is. But in software management, it is not the primary question or even the first question. In general, it is an eventual question, to be asked when, eventually, it may be worth asking. Indeed, it may be of greater value to a university, and to a research lab and its developers, and to industry that a university resist almost nearly the desire to rush toward commercial value for software, even when there is a glimmer that this may be the case. The contribution of software to a platform that becomes a standard, that is broadly acknowledged and used, may far outweigh an early opportunism to try to capture all that anticipated value in a single deal, whether with an existing company or a start up.

It’s as simple as the dog, the bone, and the reflection in the water. But who reads Aesop to learn IP management?

Software often has value quite apart from its presence as a commercial product, or as a product one that is exclusive to a single company. Software need not be sold at all. It may be shared. It need not have a product form. It may be simply a research artifact or tool, as is. And more valuable to those adopting it than if it were in a product form. It may be that software serves as an illustration for what is possible, giving impetus to multiple efforts to develop similar and better implementations. We have had instances in which companies wanted to acquire software to show their own engineers what happens if one spends any more time doing something as ill-conceived as what our lab had done. At $150K/engineer/year, it makes sense pay some modest amount to license in a university code in process just to get something bad in the company stopped in its tracks! Our benefit to the community is not always what folks think! It is to teach, not merely to demand investment or commercial products. That’s different than demanding industry turn all sow’s ears into silk purses, or state governments to provide seed funds to help us make pigs grow silk purses instead of ears. That’s different than demanding that since we work in the public interest, what we do must be better than what’s going on in other contexts.

We are contributors, not technology autocrats using IP rights to stick it to industry. Tough sell perhaps, but it’s where the value points for a huge portion of university research work actually reside. Not in commercial products. Not in spin outs. Not in exclusive patent licenses. And I’m *not* arguing that patent licensing has no place, or is done badly, or could be improved if it were not becoming so bureaukleptic and monocultured. I’m arguing that attention to software teaches a much wider range of possible engagements, in which patent work also finds itself, more truly and more strange.

It may be that having a single company as the exclusive commercial partner is the worst thing one could imagine for research software. This blog post is written with WordPress, which is not sold to me, delivered over an internet populated by hundreds of university technologies that also have not been sold, are not exclusive to any one company, and collectively form the standards by which the internet is able to operate.

Collectively, however, these technologies constitute a platform on which a huge amount of market activity takes place. Universities have a critical role in ensuring a steady flow of innovation into such platforms, whether in industry or community, whether for research purposes or as a practical matter beyond research. IP rights management that interferes with this flow of technology can do untold harm. It’s like putting a dam on a river. Yes, there may be flood control and electricity and a big reservoir for jet skis, but so much for the salmon runs, the silt essential for populating the ocean beaches, and the towns flooded out in the process. Here, it’s even worse if the IP that is taken by a central office never generates electricity or controls floods or enables jet skis, but rather just creates a drag on decisions, isolates software projects, and demands compliance to show consistency with a pre-set policy.

When has an effective norm of innovation been slavish compliance with a bureaucratic process? Only when the process itself is subject to responsive change. Otherwise, we’re barking up a tree with no squirrels.

For all of this, software has been remarkably success from a university IP point of view. In the past 15 years, the licensing deals involving software have done very well, perhaps outpacing anything in biotech in the same time. Google generated $300m+ for Stanford. Lycos at CMU. Akamai at MIT. A major share of Oregon’s work is in digital works. Not to mention all the internet, database, and research codes that have come out open source.

You won’t see this in public figures, of course, because no one in technology transfer reports only their *new business*, such as over the last five years. And since patents run now for 20 years, once you have a big hit, that income typically hides all the structure in the portfolio. Essentially, you lose a decade or more of hindsight by reporting huge success mixed with normal activity. While the big numbers may be good for politics, they are lousy for planning. Worse, if people within a university attempt to reason from the reported big numbers to planning. They won’t be able to see the last decade of portfolio until the previous decade’s work vanishes with patent expiration.

Software management pushes the portfolio toward reporting structure rather than exclusively success. For public universities, the effort then is to show what is happening, not simply skipping to put a spin on it one way or another. In this, one might start to see that even reporting a big aggregate income number may work against developing a well grounded public understanding of the activity, and that in turn can work against the effort, especially as people debate whether the public is well served by university efforts to turn research to profits.

One may point out that if technology transfer is fundamentally instruction, and licensing is finding a suitable classroom of companies, no one debates the central role of the university in hosting the activity. If however the university spins itself as a seller of research properties for exclusive corporate use, then there will be a real debate–a legitimate one–about enclosing commons, warping public values, threatening the integrity of research (one more way, at least, of doing this), and getting in the way of innovation more than helping it.

Software now shows up throughout university activity. It runs across all three forms of intellectual property. Patents show up, but are not nearly so important as copyright. And trademark may turn out to be more important than either. One might point out that few universities even have a trademark policy related to research properties (as distinct from their sports teams and university logos).

Perhaps it is time to put together a training program in software IP for university research. Information assets, really. Intangible research assets, even. Aim to make clear how software teaches a broader range of activities, of value points, of relationships between research and community, of collaboration with (and challenge to) industry and investment. Show how to use IP in this effort, conforming IP practice to the goals of software development projects rather than seeking to force software practices into pre-set patent management regimes. Then one might be on the verge of seeing how a university might migrate from a policy dominated by a well intentioned but relatively narrow view of inventions and patents to an innovation policy that recognizes foremost the university’s role in creating the pre-conditions for innovation, for economic activity, and only secondarily, from time to time, in participating in the commercialization side of things, when it really matters, where it counts for something to do so, and only with the IP needed for the initiative.

This entry was posted in IP, Projects, Technology Transfer. Bookmark the permalink.