Should software be patentable? That's the wrong question to ask

Should software be patentable? That's the wrong question to ask

Summary: Marty Goetz — the man granted the first software patent in 1968 — says a true invention implemented in hardware is equally patentable if implemented in software

TOPICS: Legal, Piracy

...sorting data on a general-purpose computer with tape drives that could read data in a forward direction, but not backwards — those were early IBM drives.

There were many sorting patents using special-purpose hardware before mine, but none that used a general-purpose computer. The following sorting patents were referenced in my patent application and were described as through and/or circuitry gates, and as methods and apparatus for sorting data: Sorting Device, Fillebrown, 5/1961 #2,985,864; Sorting Apparatus, Guerber, 5/1960 #2,935,732; Apparatus for sorting of Recorded Digital data, Dirks 3/1966 #3,242,466.

The patent was issued on 23 April, 1968, and a computer publication heralded my patent in a page 1 headline, 'First Patent is Issued for Software, Full Implications not Known Yet'. Fortune magazine called it an "unprecedented patent".

Reducing circuitry to a mathematical algorithm

Has there ever been a controversy about whether an invention using hardware circuits is patentable? I doubt it, even though hardware circuits are much like software in that they can be reduced to a mathematical algorithm.

The software products industry is competitive and needs patent protection as much as any other tech industry. Those software doomsayers who say software is just ideas, mental processes or mathematics would change their mind if they examined the different phases of the life cycle of a software product.

During the definition phase software companies describe its functionality, its specifications, the environment in which it must operate, and its operating characteristics. During the design phase, they develop and define all its interfaces, break down the functionality into modules, and do all the engineering so that the product can be properly implemented, maintained and enhanced during its lifecycle.

During the implementation phase the software is debugged, tested, and goes through quality assurance. During the delivery phase there is alpha and beta testing, documentation, installation, and training. Often software companies sell the product to other companies where the software becomes a component of a larger system and is repackaged.

During the maintenance phase the company warrants its workmanship, and guarantees the correction of errors and defects. Finally, during the enhancement phase the software is improved, enhanced, upgraded, and new models, or releases, are announced.

Note that these terms are all consistent with a manufactured product: research, competitive analysis, functionality, specifications, operational environment, operating characteristics, interfaces, modules, engineering, implemented, debugged, tested, quality assurance, alpha and beta testing, documentation, installation, training, OEM, component, system, repackaged, maintenance, warrants, workmanship, guarantees, errors, defects, improved, enhanced, upgraded, and models.

Machine processes are patentable in software

From 1968 until 1980, my previous company, Applied Data Research, filed Amicus briefs in the Prater & Wei, Benson, Johnson, Flook, and Diehr US supreme court cases, in which we argued that a machine process patentable in hardware is equally patentable in software. Here is exactly how we posed a question of law in our 1980 Diehr brief:

"Whether a computerised machine or industrial process that is patentable subject matter under 35 USC 101 when constructed with a hardware program — wired circuits — would also be patentable subject matter when constructed with a stored computer program — ie, firmware or software?"

In summary, it is a fact that software and hardware circuitry are interchangeable.

The US Patent and Trademark Office is in agreement with that question of law. In 1996 it published its Examination Guidelines for Computer-Related Inventions (Final Version) (PDF). The guidelines stated the following in its introduction:

"The guidelines alter the procedure office personnel will follow when examining applications drawn to computer-related inventions and are equally applicable to claimed inventions implemented in hardware or software."

In summary, it is a fact that software and hardware circuitry are interchangeable. As noted in Hardware/Software Tradeoffs: A General Design Principle? (PDF) the decision to put certain features in hardware and others in software is based on such factors as cost, speed, reliability and frequency of change.

So it is up to the courts and the patent office to put the same constraints and criteria on the issuance of patents that show the preferred implementation of the invention in software as it puts on the issuance of patents where the preferred implementation is hardware circuitry.

Martin Goetz was a founder in 1959 and former president of Applied Data Research, a $200m (£125m) company that was traded on the New York stock exchange before its acquisition by Ameritech in 1986. His company was the first company to sell a software product commercially and was a pioneer of the software products industry. He received the first US software patent in 1968. He testified as an expert witness for the Justice Department in the IBM suit in 1976. In 1989 he was elected to the Infomart Information Processing Hall of Fame along with Bill Gates and in February 2000 was elected to the New Jersey Inventors Hall of Fame. Currently, he is a private investor and management consultant to software product firms.

For the opposite point of view on software patents, read Software patents are legalised extortion by Mike Lee of Appsterdam. 

Get the latest technology news and analysis, blogs and reviews delivered directly to your inbox with ZDNet UK's newsletters.

Topics: Legal, Piracy

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • As a software engineering student I really don't see the need for software patents, isn't copyright enough? All the simple programs I write depend on the vast software libraries and languages written by many people much smarter than me - I'm very grateful they are openly available for me to use and learn from, I'm standing on the shoulders of giants. I just feel we would all progress faster when not bogged down in a quagmire of software patents and litigation.
  • The fact something can be implemented in both hardware and software means that any such hardware shouldn't be patentable, not that software should be patentable. Patent monopolies steal from us all. It's beyond time we abolished the disgusting patent system altogether, in all fields, not just software. It's been known for years that patents are grossly harmful to innovation and society, yet instead of dismantling the system allowing their issue, some people are pushing to extend it to more and more fields? What evil.
  • You don't get to see the source code, you only get the obfuscated machine code. This means that

    1. Trade secrets work better, because that's what most companies do now.
    2. Patents would undermine this protection because you would have to disclose the secret algo and not just the (naturally obtuse) machine code.
    3. Patents would be issued without knowledge of prior art because trade secrets work so well.
    4. If a company used patents instead of prior art, it would indicate a patent troll rather than a company genuinely seeking the better protection offered by trade secrets.
    5. If patents are granted, then you would be issuing patents to people who are not the inventors, and thus everyone would have to patent... not to protect the invention, but to protect themselves from the patent office issuing patents to later non-inventors.

    If a microchip is similarly better protected by trade secrets than patents, then it too should have patent protection removed from it so that superior trade secrets can be used.

    The question to ask is simply this, if companies are patenting to protect themselves from incompetently issued patents from the patent office, are the patents issued doing more harm than good?

    Since we started this 'give everyone patents' silliness, the USA position in the world has declined, as others have followed so their trade in the affected areas is collapsing. I think the two are linked.
  • You have incorrectly framed the debate sir, not I. Patents are meant to cover scientific inventions not protected by copyright and they represent a mutual trade between the inventor and the public. The inventor shares the "How he did it" with the public via his patent filing and the public in return grants him temporary monopoly over its use to make money back.

    This does not work with software or "business method" patents for two reasons.
    1: Software is covered by copyright so it is already protected.

    2: Business methods or the patents filed for software do not cover the "How" but only the "what" it does, and in software the "what" is utterly useless without the how.

    Imagine I wrote on a cocktail napkin the design for Doc Brown's flux capacitor. Detailed all the parts, assembly, etc. You'd say that's pretty useful right? Now imagine instead I just wrote "Delorean, flux capacitor, airplane parts, time circuits, put together, get time machine."

    That is what a business method, or "software" patent is, the latter. It's utterly useless and tells us nothing about how to build the damn thing ourselves. In the case of software though it's a double insult because near perpetual copyright terms means we will never see the "How" and because the "what" is so vague it acts as an over broad bludgeon against innovators.

    How dare you, you have misframed this debate quite grossly.
  • What you fail to miss is your own question. The question of "Is an invention that is patentable in hardware, equally patentable if implemented in software?" is answered by the question "Is software patentable?" in that it is a part of the second question.

    That is, if when answering the question "Is software patentable?" the answer is "No" (as nearly all Software Developers, Software Engineers, Programmers, etc. assert - e.g. the practitioners of the field) then the first question of "Is an invention that is patentable in hardware, equally patentable if implemented in software?" is naturally answered with a resounding "No" as well.

    If, however, when answering the question "Is software patentable?" the answer is "Yes", only then does the question "Is an invention that is patentable in hardware, equally patentable if implemented in software?" become a valid question to ask as it would help clarify the boundary of when the question "Is software patentable?" is Yes and when it is No.

    Of course, you are also completely ignoring the fact that the practitioners of the field nearly uniformly say that software is not patentable in all cases and that it is rather the non-practitioners, lawyers, business executives, etc. that want the answer to be "Yes" - e.g. the people that do not understand the field whatsoever.

    Now interestingly, there is also a big overlap of that group with the group that is trying to define software engineering using traditional engineering as a model; however, that model does not fit software engineering at all as the practitioners of the field recognize, which complicates things as the organizations do the legwork associated with software engineering only to have the practitioners ignore the legwork when implementing the software as the 99.999999% of the legwork is meaningless to them, but it makes the people that don't know anything about the field feel good that they contributed something to what they don't understand.

    In both cases, you have people that don't understand the field trying to do something for their own benefit and not the benefit of the field itself. And in both cases they are ignoring the people that really do understand the field.
  • Unfortunately we are failing to distinguish between the hardware and software. While you can argue that hardware is based on a mathematical algorithm, it is statically set and cannot be changed. Software can be changed, removed from the hardware, manipulated, reloaded; it is very dynamic. By patenting software, you are preventing others from using it as a building block, and forcing everybody to duplicate effort that has already been done. You are forcing companies to reinvent the wheel, over and over again. And the litigation resulting from software patents is out of control. Companies suing other companies, or collecting royalties, for products they don't even manufacture. Removing software patents allows multiple parties to build from one another. While this may not be what greedy companies want, services and support for software that is not protected by patents can be used for revenue generating. As an example, take Red Hat, which improves GNU/Linux with its own release, and offers pay support for its product. Users can open for the free version without support, or pay for the commercial version which includes support. This allows freedom of choice and allows a company or an individual to select which one suits best.
  • How to Fix the Patent System. Really.

    I'd prefer to abolish all business process and software patents. However, if we can't, lets tax patents like "real" property (patent lawyers often claim IP is "real" property). Just as taxes on "real" property encourage owners to develop their land, the tax would encourage individuals to develop their ideas.

    Let's say the patent office charges patent holders $1,000 per year to maintain a patent. If the patent holder doesn't like the fee, they could sell the patent, or release it to the public domain. Oh, and that tax would be a nice boost for our nation's tax revenue.

    For real innovators, it's no big deal. Its a small sum for a for a feature from which they are certainly betting a whole lot more to gain profit. For the trolls and squatters, it provides a carrying cost that will cause them to trim their portfolios significantly. And it doesn't rely on patent inspectors to have cognitive capabilities because the market would correct itself.

    This would profoundly improve the patent situation, derail Intellectual Ventures and other trolls, and realign the skills of thousand of patent lawyers to valuable contributions to society. A perfect trifecta!
  • There are two main problems: firstly, that anyone taken to court for infringement faces substantial costs even if they win the case; and secondly, that cases are held at state level, so patent trolls simply incorporate in East Texas. Fix those problems, and only then can we let the courts decide whether an invention is novel and non-obvious.
  • I agree that many have oversimplified the software patent debate; although I certainly understand the concerns of software developers and other software patent opponents, I also think there are quite valid and appropriate circumstances for such patents. Personally, I would like to see more discussion of implementing a multi-tiered patent system that granted different terms and types of protection to different types of innovations -- for instance, shorter terms and thinner rights for software patents, and longer terms and more substantial rights for pharma patents. The one-size-fits-all patent system doesn't address many contemporary needs.
  • This gentleman appears to be missing the point. The difference with something being hardware is that you can patent what the device actually *IS*, and not merely what it does, because it is a real thing. What the device does might be what makes it useful, but you don't have to patent what a useful physical thing does because you already have a useful physical invention that you can patent directly. If somebody else comes up with a sufficiently different and innovative way to accomplish the same thing, then that too ought to be patentable, and competition can reasonably occur. With software, and in particular, computer algorithms, there is no reality or substance to them at all... it is ultimately all just abstract ideas and applied mathematics... and there is no actual division between what an algorithm is and what it actually does. We can direct a computer to perform the necessary computations to invoke the effects of the algorithm, and with the appropriate devices connected to it we might even affect the real world, just as a physical invention would, but that doesn't change the fact that the algorithm is anything more than just an idea... no matter how fully thought through.

    Processes, and especially physical processes. are an entirely different beast than algorithms because they are not merely the application of mathematics or the concatenation of abstract ideas utilized in succession, like all computer software is.
  • @Michael S. Mikowski. I thought there were already ongoing payments to maintain a patent in force.

    My late father was an individual greatly ahead of his time. He had patents for which he could not afford the ongoing payments so the patents lapsed, at least that is what I have always understood. Some of these are now an integral part of our everyday experience. In particular, when I was very young he held patents on a monorail system with full working drawings and scale models. Again he could not afford to keep up the payments and the time was not right. So far as I know, these patents also lapsed.

    It seems to me that, for software at least, the US patent system is in dire need of change, but how would this be accomplished. The power is with the people who most likely either don’t want the change because they have a vested interest or have no understanding. The inertia against change is very great, as it is in countries where corruption is rife, those with the power and a vested interest will always block the change, what ever the rhetoric.

    All the ‘Mickey Mouse’ patents now granted in the US speak of gross inefficiency, inertia and lack of understanding on the part of the USPTO. We now see this spreading to the rest of the world.

    Whilst I tend to agree with some of Marty Goetz’s points, I do wonder, in the prevailing circumstances, how they could be achieved in practice. The processes have already been perverted, corrupted and dishonoured.

    In my former industry, construction, binding arbitration used to be the way to settle differences. This used to be a matter for professionals from within the industry and equitable and fair resolutions were achievable. Sadly this has now been replaced by litigation and consequently construction projects are, as a result of the introduction of the legal factor, blighted from day one.

    By the way, fantastic discussion.
    The Former Moley
  • It's gentle and spatulate to withdraw all redundant items from my machine in no second with the helpfulness of registry cleaner.