Rise of Computer Patents

It is fairly difficult to nail-down which patents are “software” patents.  For the chart below, I parsed through all utility patents issued since 1976 and looked for any of the following keywords within an independent claim: computer; processor; software; machine.readable; information; memory; data; virtual; or storage.media.


31 thoughts on “Rise of Computer Patents

  1. 11

    Dennis – I think your search terms are sufficient. However, another way to go might be to simply search for (method or process) in claim preambles within particular art units, e.g., 2600, 2400…

  2. 10

    Are “computer patents” the same as “software patents”? Computer hardware patents a traditional patents.

    “Software patents” actually is a misnomer. Conceivably the same invention may be realised both in hardware and in software. It is a means of realisation, and as such irrelevant for patent law!

    “Software patents” earned a bad reputation. But that cannot be due to the fact that they are software related. Rather bad patents are more common among (alleged) “software patents”:
    – they often describe a problem rather than a solution – so they prohibit competition without contributing to the “useful arts” whatoever
    – they often relate to (minor) parts of products, which makes them ideal for patent wars.

    1. 10.1

      If they relate to minor parts of products, why do the parties go to war over them? Why not just drop the “minor” feature?

      They do not claim the problem. They claim the solution to the problem. You are just upset about the level of specificity at which they claim the solution.

      One step of the claim reads filter x to arrive at 99.999% pure X and you think they must be limited to filtering using 4 layers of filter paper, because that is what the inventor mentioned using. But everyone knows that the filtering can be done a zillion ways and the way chosen doesn’t matter. The invention is in the combination of the filtering with the other steps….

      1. 10.1.1

        They claim ALL possible solutions to the problem. It is still just math. One solution is an identity to other solutions…

  3. 9

    Interesting data. To my eye, the graph looks sigmoidal suggesting an early quasi-exponential expansion followed by dampening. The inflection point seems to be about 2008. It may be that the Great Recession had an impact on what would have otherwise been exponential growth. Or, it may be there is a natural saturation of increase in computer patents. Or, it may be that I’m seeing a shape in the data that simply doesn’t exist because sigmoidal curves are among my favorite shaped curves.

    1. 9.1

      Stuart – one feature of the data here is that it is shown as a proportion of all patents. That means it has hard limits of 0% and 100% and, as a consequence, exponential growth is unlikely. A graph of the raw numbers very nicely fits an exponential curve.

      1. 9.1.1

        Wow (the following comment intended only with levity)

        The Amish are really going to be p1ssed with that curve.

  4. 8

    Perhaps add programming and coding

    Punch cards and even relay flips were maybe not fully software yet they are logically equivalent to what we call software, and cards are machine readable, relays create state change. Early computers were limited by the available technology rather than by conceptions of usefulness.

    MIT’s Whirlwind computer from 1951 in marketing literature touts itself “a general purpose electronic computer” listing uses: “Air Traffic Control, Industrial Process Control, Gun (military ordinance) Control, Insurance Handling, Mathematics, Inventory, Economic Analysis, Simulation, Census and Scientific and Engineering Computations” The flow charts still look very familiar.

    You who believe all is new are now closing your eyes to the 1952 era states of these arts. The PTO competency could use technology historians aware of foundational books available that address lack of institutional knowledge.

    1. 8.1

      “Perhaps add programming and coding.”

      Or perhaps not. Dennis has searched only independent claims here, in an attempt to identify the subject matter of the invention itself (and not, e.g., an incidental reference to implementation details, prior/related art, or alternatives in the description).

      “Programming” and “coding” are not words that are likely to appear in patent claims.

    1. 6.1

      Good! I was wondering what new insights this (draft) post was aiming for. Obviously the significance of the computer industry and its core technologies has grown exponentially since 1976. My guess is that similar curves would result if you plotted the occurrence of the same search terms in the wall st journal, PhD theses and academic publications, course catalogs, resumes, job postings, product advertisements, etc.

      The surprise would be if the curve for patents did *not* show a steep rise in occurrence of these terms over this period.

      Maybe if you generated curves using the wall st journal or nytimes as a kind of baseline for comparison, we could assess whether the rise in usage of computer terms was especially steep (or not) in patents.

  5. 5

    As you get back into the 70s and even the 80s, the “software” used by a computer can have quite a different form than “software” does today. For example, punch cards were still being used as the “software” in many computing system well into the 1980s.

    1. 5.1

      I remember using punch cards in the early 1980s. And Fortran. I personally didn’t consider punch cards to be “software”, as they were heck to change.

      1. 5.1.1

        You don’t think you could write an e-commerce platform on punchcards using fortran? MM can.

  6. 4

    By 1976, stored program electronic computers (and the recognition of their nearly unlimited utility for assisting with information-processing tasks) was at least three decades old.

    And yet the PTO, with the help of some fantastically short-sighted jurists, spent nearly four additional decades pretending that programmable computers were invented yesterday because that’s how their “stakeholders” liked it.

    Now we’ll find out what happens when the USPTO is forced to rethink its longstanding practices. If people thought the “reject reject reject” era was scandalous (it wasn’t) then those same people are going to lose their marbles completely if these yearly grant rates don’t start plummeting. That’s because if the grant rates don’t start plummeting, we can be quite confident that the PTO is corrupt to the core and in need of some housecleaning from top to bottom.

    1. 4.1

      I don’t think you should be so quick to judge corruptness when incompetence adequately explains the situation.

      That said however, since the PTO staff is probably the bulk of the most competent people in the area in the country if they are likewise incompetent there may simply be no better solution than them.

    2. 4.2

      1946? Are you speaking of ENIAC, the machine that was as large as a house? And consumed 150KW of power? It’s a bit hard to say that “electronic computers” were three decades old as of 1976, when there were so few of them.

      Regardless, a computer that is programmed to perform something new and useful should be patentable.

    3. 4.3

      “Corrupt” is not cabined by your Luddite understanding of 1976 tech. You old guys who can’t understand your latest I-phone seriously should consider retirement. I have two words for you, Festus: “Sun City”.

    4. 4.4

      MM, may I suggest the real problem with the CCPA/CAFC stemmed from the reaction of many in the patent bar to the Supreme Court in the ’40s – that lead to the some changes in the ’52 Act intended to coral that court. But the real kicker was the appointment of patent bar mover and shaker Rich to the court — who had the singular agenda to make everything statutorily based so that the courts could no longer do their own thing — and this included channeling all patentability decisions into 102/103, his baby.

      His instincts to confine patent law to the statutes was, in my view, proper. But his further instincts to ignore precedents, especially the Supreme Court precedents, was wrong.

      Thus his thinking on 101 was to ignore it, and force everything into 102/103, and that thinking seems still to be the PC thinking of the patent bar who as a whole seem truly dismayed by Alice.

      On programmed computers, the CCPA went off the rails with Bernhart and Benson. That the bar, and the CCPA/CAFC, could seriously consider these good precedents after Gottschalk v. Benson is symptomatic of the overall problem of the CCPA/CAFC in the Rich era. Hopefully, with Alice, that era is gone in the courts.

      But you a right, the PTO needs to get squared away and should stop issuing patents as if Alice did not exist.

  7. 2

    That’s tough to parse. I do a lot of control systems for machines, and all of those patents will include some of those terms in the boilerplate.

    1. 2.1

      Boilerplate for the independent claims? I narrowed it only to those because it is the area most likely to be devoid of fluff that is not really part of the invention.

      1. 2.1.1

        Oops, I missed the tidbit about independent claims. No, of course not. The only term that would show up in the independent claims for these types of applications is controller.


          Strike that – data and information definitely appear in a lot of these claims because that’s the physical sensor data or information fed into the controller.

    1. 1.1

      No, all of the effort that goes into programming software is just doing the obvious. We need Chief Justice Roberts to start programming some software, and I’m sure you would see how he can make a single program that performs all of the obvious stuff that millions of programmers have been scamming their employers with for decades.

      1. 1.1.1

        Thats what the 4G languages were supposed to do…

        The end result is a “point-and-click” programmer that doesn’t really understand what he is doing…

Comments are closed.