The Patent Exchange

Thu, Aug 11, 2011


Software patents have been a hot topic in the Apple community over the past few months. Not only are there a seemingly endless stream of trolls coming out of the woodwork in attempts to extract non-deserving riches from folks who actually produce something, but Apple is currently embroiled in a number of high profile lawsuits with respect to iOS design and technology.

As a result, some have taken to the virtual streets and declared that the patent system is broken.

Nilay Patel of This is My Next disagrees and writes an impassioned piece detailing the function of the patent system and the challenges involved in preventing widespread abuse.

While the patent system is largely viewed as a mechanism to encourage innovation, Patel writes that the dissemination of information is also part of the patent system’s foundation:

The core public policy behind the patent system is widely ignored, even though it’s extremely simple and really quite clever. Patents are more than just a simple incentive for people to develop new inventions — they’re actually an exchange between inventors and the public. In exchange for a time-limited monopoly on their inventions, inventors must fully disclose the invention itself in the patent specification, and agree to release their work into the public domain once their monopoly runs out. The rules for disclosure are laid out in 35 U.S.C. § 112, and they’re fairly strict: the specification must be detailed enough so that anyone with “ordinary skill in the art” of the invention can build the claimed technology, and they must also disclose the “best mode” of building the invention. Breaking the rules can have severe consequences, since a patent that doesn’t adequately disclose the claimed invention can be ruled invalid. And since patent specifications fall into the public domain once the patent expires, we get a huge and constantly-growing vault of fully-disclosed technology that anyone can use to build new products.

Patel’s article is a long read but well worth it.

That’s not to say that there aren’t valid counterpoints to many of his assertions.

To wit, a commenter on Hacker News chimes in:

Nilay, I think your patent exchange misses a key point in software. No one reads them. Nobody reads patents. In fact, go talk to a Microsoft and Apple engineer about patents and they’ll tell you that not only do they not read them, their corporate policy does NOT allow engineers to read patents. Apparently it drastically increases the likelihood of treble damages for the company.

You give the PageRank example, missing though that there was work from IBM that was very similar. See this classic paper, “Authorative Sources in a HyperLinked Environment” from 1997(…. The genius/luck of Google was continuing to push this idea of search, when no one else seemed all that interested (recall AltaVista, Yahoo, and Lycos had all seen Google’s results and were offered a chance to buy the company — they all passed. Their genius was in persisting and not giving up and getting their PhD.)

The Apple patent you show is an example of what I call, “Being the first to ask the question”. I’m not sure how else you’d solve the problem besides how they did it. It’s the obvious way to do it. They probably lucked out because they were the first company to be faced with the question. This has become rampant in the mobile industry. Whenever you have a new form factor, there are new problems. They aren’t necessarily hard, but they’re new. And the fact that they’re new problems means there’s no prior art. You can suddenly file a bunch of patents based on your solutions, 99% of whcih are the same solutions the guy across the street would come up with in six months when he happens to hit the same problem. That’s not innovation — that’s blocking innovation as it ensures your six month advantage becomes a 15 year advantage (or however long patents expire).

Nobody in SW reads patents. Furthermore patents are actually very hard for those in the field to follow. They use non-standard jargon. I’d much rather read source code or a CS paper to get the ideas — as those are usually written just by the actual developer, and not translated by the lawyer.

And an unrelated, but important point, IMO. SW, unlike most other endeavors, is something that people rapidly build on. People still take aspirin today, in the exact same form as 50 years ago. No one uses a piece of software in the exact same form as 50 years ago. The closest is probably vi, but even vi has had significant code churn over the past 30 years — it’s quite possible that it currently shares no lines of code from the original version. SW evolves rapidly. Patents seem much better suited for fields where inventions can stand on their own for significant periods of time.


1 Comments For This Post

  1. Krabbie Says:

    Edison made it to the patent office JUST before “whatshisname?” who you don’t re
    Ember. So, your point is? Getting there IS the salient issue in patents. IMHO

eXTReMe Tracker