Patentability of Softwares: More Canon Fodder

Patentability of softwares, it appears, has presented the academic community the opportunity of a lifetime to produce tomes and tomes of literature, quoting the same set of judgments over and over again forgetting that brevity has its own benefits and attraction. The latest judgment that has stirred up debates on blogs the world over is the Symbian Ltd. v. Comptroller General of Patents given by the High Court of Chancery Division in the UK. On a completely different and irrelevant note, I never understood the reason for the extra “p” in Comptroller and decided to look for its origins. As always, Jimmy Wales’s creation gave me the information I needed; those interested can check this out. Anyways, returning to the discussion, before I could form an opinion on the judgment, I had to read two other judgments to make sense of the controversy. The confusion that the Symbian judgment of the UK High Court of Chancery Division has supposedly caused is with respect to the applicability of Aerotel/Macrossan test laid down in the Aerotel Ltd v. Telco Holdings Ltd case and apparently is at loggerheads with European Patent Office (EPO) Board of Appeals rulings on software applications, especially the Vicom case. So I read the Aerotel case and the Vicom case to understand the issues in the Symbian application.
I am sure most of us are familiar with the Biblical portrayal of God as this well-built old man with flowing white beard and a voice which is at once filial and majestic. When I read the Aerotel judgment, I pictured Justice Jacob in exactly the same way for it was nothing short of a pleasure reading the judgment to the extent that I imagined myself actually listening to him deliver the judgment (remember Harry Potter being sucked in Tom Marvolo Riddle’s diary?) and I never realized that I had read 41 pages covering the entire set of case laws on the issue. The four step test of the Aerotel case has helped reduce the confusion that software-related patent applications create in the UK Patent office. The test, as most know, is a reformulation of the test laid down in the Fujitsu case and is as follows:
1. The claims in the patent are to be properly construed to understand the scope of monopoly sought
2. The actual contribution made by the inventor has to be identified
3. Next it has to be determined if such contribution falls solely within excluded subject matter and
4. Finally, one has to verify if this contribution is technical in nature.
This test has been guiding the UK patent office for sometime now; also, office practice notices issued in November 2006 and February 2008 refer to the test. There has been a certain amount of disagreement on the application of steps 3 and 4 of the test with one group saying that once a determination has been made in step 3, step 4 would merely serve to affirm such determination and so is inherent in step3 with no individual standing. The other group differs and feels that this is not so in all cases since step 4 serves as verification and not merely affirmation. I am of the opinion that step 4 would be largely answered once one is past step 3.
Moving to the issues in the Symbian case, the application, in brief involves, the use of an interface between the executable program and the dynamic library link (DLL) in mobile phone applications with the DLL as applied in the Symbian application having two divisions as opposed to one in prior art. The DLL is a collection of small programs and files which have applications common to several functionalities of the phone. The DLL responds whenever any function is called by the user. In the prior art, this DLL either worked on link-by-name system or link-by-ordinal system. In the link-by-name system, the DLL identifies the function by its name from a table which contains the name of the function and its ordinal number (position). In the link-by-ordinal system, DLL identifies the function by its ordinal number which makes this method easier and faster since the memory used is less due to absence of names. However, upgradation or addition of new functions has its own problems since sometimes the ordinal number does not respond to the user’s call or it responds incorrectly. For a better explanation of the technology, a reading of the patent or the technology in general is suggested. The Symbian patent seeks to overcome this problem by dividing the DLL into 2 groups, one group with a fixed set of functions which cannot be changed and the other with optional functions which can be upgraded. The interface, which is another computer program, contains a mapping of all the functions and its ordinal numbers which in case of an upgradation, links up with the functions in that division of DLL which has optional functions. This reduces the confusion with respect to the ordinal numbers and increases the reliability of the system.
All the techie talk aside, in simple terms, the invention claims this interface which helps reduce malfunction of the DLL in addition to an apparatus which employs this interface and thus increases reliability. The question, as usual, is if it falls within the definition of patentable subject matter under s.1 of the 1977 Patents Act of the UK? Since all European Courts try and conform to the EPO decisions and the European Patent Convention (EPC), naturally the discussion would veer towards Art.52 (2), Art.52 (3) and Art.53 of the EPC. Accordingly, does the invention essentially and substantially claim a computer program? If yes, a patent grant is not warranted. The hearing officer in the UKIPO held that the Symbian application does not fall within patentable subject matter applying the Aerotel test. I personally agree with her for the following reasons.
Step 1: First, in this case, the claim is for the method which applies the interface and an apparatus which employs this method.
Step 2: The actual contribution is the division of the DLL into 2 parts and the use of an interface to better coordinate the mapping of the functions.
Step 3: If the invention’s heart and soul is a computer program with no hardware interface, then such an “invention” is not an invention under the 1977 Act or the EPC. This means the invention shall not be about a computer program as such or in Indian terms, it shall not be a computer programme per se. There has to be something more which lends to the invention the requisite character that would make it patentable. What is that “something” and more importantly, what is the de minimus line for that something which will just bring the invention within the bounds of patentability? To understand this one needs to know the reason for the confusion surrounding patents on software-related applications. The argument against grant of patents to softwares is that the line distinguishing an abstract concept or a principle from a concrete technology in softwares is blurred and in most cases, it tilts towards the former. This makes it imperative for the inventor to build a system surrounding the software ensuring the presence of a technology and not a mere idea. This is crudely translated as the requirement for a hardware interface since it is assumed that the presence of a hardware interface points to the “technical” nature of the problem. In other words, the burden of proving novelty and non-obviousness is shifted to the system which comprises of the software and the hardware part, for such questions are absolutely irrelevant in the context of a software program alone since softwares per se are outside the scope of patentability.

This means that if one uses a computer program in a known piece of hardware, such system may be granted a patent though the novel part is the program which falls within excluded matter, if the use of the software enhances the performance of the hardware. To put it in another way, the Act exhorts the inventor to try and divert the attention of the examiner to the system though it involves the use of a computer program. This is what the Aerotel case lays down in effect. However, the High Court has erred in its application partly because it has even misunderstood the Vicom case and the facts in the light of which that judgment was delivered. On reading Paragraph 38 of the Symbian judgment, one will understand that the Court has put the carriage before the horses. I quote the Court here:

Suppose a program written for a computer that enables an existing computer to process data in a new way and so produce a beneficial effect, such as increased speed, or more rapid display of information, or a new type of display of information. It is difficult to say that these are not technical effects, and, indeed, that the programmed computer, itself a machine that ex hypothesi has never existed before, is itself a technical article and so in principle the subject of patent protection.”

If one where to extend the Court’s logic, every new software changes the function of the computer in someway or the other for the better, so would such a computer become patentable? In effect, isn’t it the program as such which is responsible for the change in the functioning of the computer? How can technical effect alone decide the patentability of the subject matter? Agreed that patentable subject matter must necessarily produce a tangible technical effect, but merely because a technical effect is produced, that wouldn’t bring excluded matter within the realms of patentability. The Court has jumped to the issue of technical effect without first clarifying as to what causes that technical effect. In this case, the technical effect referred to is the increased reliability of the system. Applying the Aerotel test, one is expected to first enquire the reason for the technical effect. Further, the Court has gone on to support its logic using the Vicom decision saying that if increased speed was the technical effect which lent patentability to the Vicom invention, why shouldn’t increased reliability do the trick for Symbian? The logical explanation to the Court’s flawed parallels is its ignorance of the facts in the Vicom case. In the Vicom patent, a set of claims dealt with the use of dedicated filters to apply the method claimed in the patent thereby satisfying the system requirement and also the technical effect requirement. This is not so in the Symbian case. The UKIPO has decided to appeal against this decision and hopefully, better sense prevails in the Court of Appeal.


  1. AvatarSushant

    The article was really confusing. I did not understand what laws you were referring to and what all judgments that you are citing. After searching on the web I realized that you are talking about England.

  2. AvatarJ. Sai Deepak

    Dear Sushant,
    I apologise for the confusion caused. I have modified the article to mention that it is a UK judgment which is being referred to here. I hope the confusion is restricted to the jurisdiction, kindly let me know if the analysis too was confusing.

    J.Sai Deepak.

  3. AvatarSushant

    I got confused because I was thinking that you are talking about Indian law. May be I read this in haste.

    I have never programmed on a windows environment but it seems strange that the linker is trying to resolve function names by “ordinal” instead of name. AFAIK, the linker ld on GNU/Linux environment resolves based on function names. I wonder how much faster is resolution by name when compared to resolution by ordinals. If you are creating a binary search tree over ordinals for faster search, then you will be consuming more memory.

    But beside that point it just appears weird that you can get patent for such stuff. You correctly point that out but I do not know how you can argue such things in law.

  4. AvatarAnonymous

    “In T 125/01 the board held that a pure software enhancement with no changes to the hardware could constitute an invention within the meaning of the EPC if the modified software made a technical contribution to the invention and therefore could not be isolated as such. The invention concerned a module for controlling a communication device, preferably a car radio, and differed from the closest prior art only in respect of the program loaded, which was intended to make the control module easier to adapt and thus flexible in use, with a view to allowing additional technical device functions to be added and/or making the device easier to operate. In the board’s view, even the problem description showed technical aspects, as it would certainly be viewed as technical in conjunc-tion with a hardware interface. The board also felt that the appropriate restructuring of a control program to that end also made a technical contribution to the invention, by analogy with the redesign of a hardware control device (eg in terms of electrical circuits or mechanical elements).”

  5. AvatarJ. Sai Deepak

    Hi there,
    I see your point, but I am not sure I agree with the parallels you seek to draw between the Symbian application and the T 125/01 case. If I have understood you correctly, your point is that so long as the use of software solves a technical problem, notwithstanding the fact that it uses existing hardware to give a better solution, the product constitutes a patentable invention. Though this is not entirely inaccurate, I believe that there is more to the problem than this. Let me explain this with the help of an analogy.

    If an inventor comes up with a new nucleotide sequence, what he has done is to merely discover the sequence and he will not get a patent for such a discovery. But if he decides to apply this discovery and use this nucleotide sequence in an existing cloning vector, the product is a patentable invention. Now, here if one were to blindly separate the nucleotide sequence from the cloning vector and say that since the sequence is a discovery and the vector is prior art, the combination is not patentable since the individual components which go on to make up the product are not patentable in themselves, such an argument has no merit. The reason why I say this is because ultimately every invention is nothing but a contextual application of a scientific principle to solve a particular problem and the law requires such problem to be technical in nature. This application of the principle or discovery in a very specific form is what we call technology.

    It might appear that the example of the cloning vector which I have given supports your argument, which is the patentability of softwares that solve a technical problem using existing hardware. Unfortunately, it is not so for one needs to know the reason for patent law’s obsession with a technical effect to understand the controversy surrounding computer-related inventions (crudely and inaccurately called software patents). I strongly believe that patent jurisprudence has been heavily influenced by mechanical inventions and this was the case for a very long time until pharma and software inventions started shaping the course of patent law. However, the influence of mechanical inventions still lingers and the idea and concept of a technical effect is a product of this. The change brought about by mechanical inventions in most cases is very visual in nature and its effects are there for everyone to see. This is one reason that we still expect such visual changes or “technical effects” from other technologies as well when such an expectation is not entirely fair to the technology. The nature of the technical effect varies from one field to another. However, as the law stands it requires a kind of visual confirmation to show that the technology has improved the prior art. Kindly refer to T 0115/85 and T 469/03.

    Having said this, it must have become clear by now that computer-implemented business methods do not satisfy this requirement and so are strictly out of bounds for the problem they solve is not technical in nature. Even if a software solves a “technical problem” as such, there have been umpteen cases where patents have been denied. If what you say is correct, there exist very few situations where softwares do not solve a technical problem, the how do you explain the denial of patents for such softwares? I shall try to rationalise this. Let us take the example of a particular manufacturing process which previously required the presence of an operator at the control console. If this entire process is automated using a computer, with the system being designed to respond to the computer, will such a system be patentable? In 9 out of 10 cases, the answer is a yes. This is because the computer here performs what was earlier the job of an operator and the use of computer brings with it certain natural advantages such as more diligence, less fatigue etc. To put it clearly, the prior art was strictly not in the field of computers and the use of computer to get a better result in an existing form of manufacture or any other process adds value to the system. So though the use of the computer is the reason for the improved performance, the claim is not for the software as such. Also, here the problem is technical in nature. In another example, if the prior art was a kind of a computer game and the improvement is to the features of the game by the use of a software, the claim will in essence speak of a software as such. In such cases, I do not see the possibility of a patent grant by the EPO or the UKIPO. Applying the logic of my first example, the decision in T125/01 which you have cited is justified considering the facts of the case and I do not see any inconsistency in my arguments and the decision.

    Let me also go a step further and state that I do not see any reason why softwares per se should be kept out of the patent regime? Is it because that the conceptual transformation from a mathematical principle to a programme is too close making it essentially a scientific principle? This may not be true in all cases and I believe that just as in any other technology, this decision should be left for a person skilled in the art which is anyway notional. However, this is for those in power to look into. Kindly let me know if you still feel that my logic is incomplete or gross.

  6. AvatarAnonymous

    I agree with you on that an was trying to only put forward the EU position.

    You shall find a whole portion dedicated to computer-implemented inventions in a case law compilation on the EPO website.

  7. AvatarAnonymous

    actually i do recommend if possible- looking at the section on computer implemented inventions. Shall email it to you. your point about why keep softwares per se out of patentability – I think EU has a more reasoned approach towards them and not a blanket rule. However, i do not agree with your visual effect concept for deciding technical nature of a problem etc.

  8. AvatarJ. Sai Deepak

    I shall wait for the email. Also, the visual effect concept is not my concept, all i said was that it has featured in some of the EU decisions and that the concept may not be applicable in all circumstances. It must be understood that in cases involving sotwares or any technology for that matter, one needs to first judge the merits of the technology in itself before beginning to regurgigate the case laws.Half the time, the discussion on the technology in hand is restricted to just a couple of pages and the dicussion on case laws runs for the rest of the judgment. Since technology is essentially a matter of fact, unless we are clear on that, to force fit the concept purely in terms of law is to turn a blind eye to the fact that patent law differs from other branches of law. To prove this, i suggest that you try and build a fluid logical structure on the basis of these EU decisions and i am sure the process will not be as fluid as it appears.


Leave a Reply

Your email address will not be published.