Microsoft wants patent system fixed (but not eliminated)

Microsoft wants patent system fixed (but not eliminated)

Summary: If you've followed any of my coverage of the patent mess that the technology industry is in -- and it's a real mess -- you'll know that I'm in agreement with anyone who says it's a mess but not necessarily in agreement with those who say get rid of it, or those who say to keep it.

SHARE:
TOPICS: Patents
9

If you've followed any of my coverage of the patent mess that the technology industry is in -- and it's a real mess -- you'll know that I'm in agreement with anyone who says it's a mess but not necessarily in agreement with those who say get rid of it, or those who say to keep it.

Topic: Patents

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

Talkback

9 comments
Log in or register to join the discussion
  • Privatise the patent office

    When I buy a house from an advocaat, he does all the necessary legal checks to ensure the ownership and stability of the house. He stands behind his opinion and can be sued for considerable sums of money if he is wrong.

    In the same way the Patent office should be made to stand behind its patents. If they issue a patent, they are asserting that: "this is a new and novel and non obvious invention". If the court decides this is untrue, then the same rules should apply, with costs and damages brought against the patent office for its false statement.

    But you should go further, privatise the patent office and take away their monopoly. Any organisation that meets the basic criteria to issue patents and has sufficient capital or insurance to cover the claims should be allowed to issue patents.

    In each case the body needs to be independant of the entities it issues patents for, of good standing and will be made to stand behind the patents it issues.

    That will bring the patent office into line with other professional bodies, solicitors, advocaats, structural engineers, etc.

    I think the recent schenanigans with the EU patent office has shown just how out of control they are. They issued software patents without a legal basis, then slotting in a loophole into a directive to create a legal basis *afterwards*. It seems to point to empire building by civil servants.

    The problem is that it is in the interests of the civil servants to issue ever more patents in ever more areas regardless of the damage they are doing to industry, so they can build their empires.

    I think there needs to be a push back against the patent office and this would provide it.
    Nigel Johnstone
    • Developing the idea further

      Company X wants a patent, Organisation Y issues the patents and asserts that it is a novel, new, non-obvious invention.

      Company X tries to enforce its patent and the court decides the patent is invalid.

      Organisation Y is now liable for the full costs of the court challenge and damages relating to that patent because they asserted that it was new novel and valid, essentially making a false statement.

      i.e. instead of *partial* liability, they should be *fully* liable and the companies applying for patents would pay in terms of their fees to cover the liability insurance.

      I also figure it has other benefits, for different industries there are different ways to determine prior art. Each patent issuing authority would have an incentive to develop ways of determining prior art, to reduce their liability costs.

      This (for me) is the big problem with pure software algorithm patents, you cannot determine prior art if the prior art is in closed source software. To make it much worse, there is an incentive for a patent office to minimize prior art: the less prior art, the more patents they can issue, the more people want to patent, the bigger the patent office.

      So the situation as it is now, can only get worse because the incentives are there for the patent officers to make it worse. Again introducing competition and liability into the patenting process would fix this.

      Imagine a system where to obtain a patent for an algorithm hidden in software is very expensive because prior art cannot be determined and hence risk is high, but a patent on software that interfaces to the outside world is cheaper because it is visible and prior art can be determined much more easily.

      Perhaps it would also resolve similar but different problems currently facing patenting of genes, and in future nano-technology which again will need different ways of determining prior art.
      Nigel Johnstone
      • I can't agree

        Liability costs big money.
        While the private sector generally out performs the public sector, throwing profit and liability into the mix would make patents impossible for anybody but large corporations.
        This is in effect a government service and must be available to Joe Blow if he has an idea.
        In theory he can do it himself but won't if he risks losing his house.
        Using the court system as part of the patent system would help bring the whole thing to its knees which mightn't be a bad thing.

        Joe
        seosamh_z
        • But thats the point

          "Using the court system as part of the patent system "...

          But thats what happens now. The only difference is that when a court says a patent is invalid and shouldn't have been issued, the patent office has no comeback for its mistake.

          Worse still, they have an incentive to issue bad patents because they have a challenge process that costs money to perform too. (You can challenge a patent but have to pay the patent office a second time to do it!).

          I think that will lead to more caution on the part of patent issuers (since they have to stand behind their patents), and so less flaky patents and fewer lawsuits.

          I also think the price for issuing patents is artificially high and competition could reduce that to bring it more into line with other professional services like structural engineering.
          Nigel Johnstone
    • Can't agree

      As lax as the process for issuing patents is, I can't imagine that handing the USPO's functions to private agencies would help matters. If anything, private patenting agencies would have even less of an incentive to run the system honestly than the bureaucrats that run it now; after all, the USPO takes in the same amount of money whether a patent is granted or denied; the same would not be true of private agencies in competition with each other.

      Furthermore, it just doesn't seem right to me to put private agencies in charge of issuing legally enforcible monopolies, any more than it would make sense to have them issue drivers licenses or building permits. Granting of legal privileges is one of the core functions of government, ranking up there with defending the country and maintaining public order (no, we should not have private security agencies policing the streets). One can argue whether patents should be granted at all, but if they are, it should be done by civil servants insulated from politics and acting on behalf of the public, rather than patent applicants.

      A better approach would be to give the USPO the money it needs to hire more and better patent examiners. If the law doesn't require it already, the applicant should be required to provide a working implementation of the invention which would either be shipped to the patent office at the inventor's expense (both ways), or if that is not feasible, the inventor would pay for the privilege of having an examiner inspect it on site (don't think the taxpayers should bear the cost, do you?). Software patents, if they're granted at all (I don't think they should be allowed) should require a working implementation, with accompanying source code and documentation. Similar rules should be established for drug and biological patents, to the extent they're permitted. Application fees should be high enough to cover the actual cost of investigation (again, the taxpayers have enough to pay for already).

      If we're going to have a patent system, we need to have one that's run honestly and in the public interest. If the rules cannot be effectively enforced, then either the scope of patents needs to be scaled back or the whole system needs to be abolished.
      John L. Ries
      • Isn't a patent just a 'statement'?

        "after all, the USPO takes in the same amount of money whether a patent is granted or denied; the same would not be true of private agencies in competition with each other."

        Indeed you have to have liability for overturned patents or it wouldn't work. If you think about it, its a few thousand dollars for a patent and on average $3million to defend against a bad patent, so the benefit is small but the damages potentially big.

        So there is no way an organisation would knowingly issue bad patents and expose themselves to those penalties willingly.

        Thus I believe it will come down to judgement and effort in determining prior art that will determine who succeeds and who fails in patent issuing.

        "If the law doesn't require it already, the applicant should be required to provide a working implementation of the invention"

        I used to believe this would fix software, but then I realised that both the invention and any prior art has to be disclosed for the patent to be determinable. A working implementation does not mean the patent can't be struck down. With software, prior art (to determine novelty and non-obviousness) is nearly always the patent killer.

        For me it comes down to a simple idea: A patent is just a statement from the patent office: "We believe invention X is new novel and non-obvious."

        They make the statement, they accept a fee for making it, but won't stand behind that statement.
        Yet if I get a structural survey the surveyor is required by law to stand behind his numbers, my accountant is liable for errors etc. Notary publics are liable, Advocaats,...

        I would have more faith in an Electrical engineering patent issued by the IEEE, or a software patent issued by the BCS. Especially if they are required to defend their opinion and pay if they lose.

        To me, the current situation is this:
        Companies pay billions in costs associated with bad patents. There are clear problems with prior art which we discuss at great length, but the patent office does nothing to fix these because there is no liability to them for their errors.
        Nigel Johnstone
        • Theres a few other plus's

          Suppose we have 10 patenting agencies. (All of good standing, independant and with sufficient capital or insurance to defend lawsuits, the usual conditions).

          A) The 'Dr Nick' patenting office issues patents easily but thus accepts more bad patents.

          B) The 'IEEE' patenting office takes much more care but only the best, most watertight, fully working patents can be issued by them.

          The other 8 are between these two extremes.

          Obviously DrNick would accumulate the bad patents over time, because he would have to charge more to counter the lawsuits, and so only the inventors with poor ideas would have to pay the extra money to get their patent.

          +1) This is a plus, bad patents get grouped into bad patent authorities with lower reputations and are thus easier to defeat in court.

          Now suppose 'Dr Nicks' patenting office is just tooooo bad at issuing patents and he goes out of business. What happens to the patents he's issued?

          Simple, they can no longer be enforced unless another patent issuer is prepared to back them.

          +2) So if a patent office is extremely bad at issuing patents they would cease trading and hence a whole group of their bad patents would be invalidated in one swoop. That would be another plus.

          But what would happen to a good patent in DrNicks patent authority if there is one, when DrNick goes out of business? Easy, the inventor can ask any other patent issuer to back the patent. If its a good patent they will back it for their fee, if not his will sink with DrNick.

          -------

          Back to the idea of free choice of price in patents. Suppose a high risk patent (one in a field where prior art is extremely difficult to determine) costs more than a low risk patent field.

          I think that 'non-obviousness' is a difficult thing to determine, what I consider non-obvious may be completely different to a judge or a patent officer. So its not a sharp cut off, its a smooth transition.

          i.e. to me Microsofts IsNot patent application is (&a!=&b) and screamingly obvious.

          To a patent officer perhaps not.

          To a judge I would spend a day explaining it and perhaps I would succeed, perhaps not.

          The variable price would cover this, patent offices would use their professional judgement to assess whether they believe it is obvious.
          So the less novel and new a patent, the more it costs and the higher the risk to the 'inventor' in applying for it. (Inventor in quotes because if it is obvious then its not an invention and he is not an inventor).

          +3) So speculators pay a penalty by paying more for their patents than true inventors of novel inventions.
          Nigel Johnstone
        • Oh I missed some, +4, +5

          When the patent office issues a patent and a company tries to use the patent against someone, the patent is often challenged in court. That situation exists because the patent office make so many mistakes, they have no reputation and their patent has not standing. People challenge patents because there's a pretty good chance they will win.

          +4) The IEEE patent issuer (in the previous post) almost always wins its patent challenges because it take more care, so when faced with an IEEE issued patent companies will be less likely to try to defeat it in court. It would be wasted money!
          -> Fewer court challenges.

          +5) So an inventor who can get an IEEE patent only pays the IEEE fee, not the patent fee + millions to pay for the patent challenges. Lowering the price of enforcement (as well as the price of issue) to the true inventor.
          Nigel Johnstone
      • Re: Can't agree

        "Granting of legal privileges is one of the core functions of government, ranking up there with defending the country and maintaining public order (no, we should not have private security agencies policing the streets)."

        Given the Founders distaing for standing armies, they would be appaled at the current police state. They intended for private citizens to handle most law enforcement, not full time giovernment employies. Don't forget the governemnts job is not to protect individuals, but society as a whole. What about protecting the individuals?

        "If the law doesn't require it already, the applicant should be required to provide a working implementation of the invention"

        That used to be required. It was later removed due to storage problems.
        STDog