On 30th August 2024, the DHC gave out two decisions on appeal against the rejection of two Patent applications filed by Blackberry Limited. In this post, I will refer to the two decisions as Blackberry I and Blackberry II. Here, I will contrast the two judgements and answer whether the reasoning employed in them is consistent and reconcilable. I also mention the Mayo Test from the U.S. to understand how it applies in the Indian context, specifically 3(k), and whether the Indian Courts, knowingly or not, have aligned their approach with that test.
Blackberry I: Purely related to Algorithm
In Blackberry I, the invention, which allowed easier configuration of data between wireless servers, was rejected by the Patent Office as a ‘computer program per se’ u/s. 3(k). The Order said that the invention was a ‘set of instructions’ which was purely functional and without any inventive hardware.
The question before the DHC was simply- Is the invention non-patentable u/s. 3(k)?
While the Court acknowledged that the invention exhibited technical contribution (conflict resolution, data management strategies etc.), nevertheless, it refused to grant the Patent since the contribution was a result of a ‘sequence of instructions and procedural steps.’ It concluded in Para 41, that since the ‘core functionality’ of the system is driven by a complex sequence of instructions, characteristic of an algorithm, it is not patentable u/s. 3(k).
Computer programs, generally, are comprised of a set of instructions which are algorithmic in nature i.e. a set of instructions. (see Sec. 2(ff) of Copyright Act) As I have explained here and here, if the computer program, despite consisting of algorithmic instructions, has a technical effect or contribution, it is patentable u/s. 3(k). The basis for the test of ‘technical effect’ is found in the language of 3(k)- a mathematical or business method or a computer program per se or algorithms. Since only ‘computer program’ is qualified with the words per se, the Court’s reasoning goes, that if the invention is something which goes beyond the computer program i.e. it has a technical contribution, it is patentable. (here) On the other hand, if the invention relates to, let’s say algorithm, the bar to patentability is absolute, and the Court won’t go into the question of technical contribution. (here)
In the present case, the Court had acknowledged that the invention did exhibit a ‘technical contribution.’(see Para 58) However, it classified the invention as an ‘algorithm’ rather than a ‘computer program per se’ u/s. 3(k).
To become eligible for a patent, the Court suggested a way in Para 52- “if the algorithm instructions are implemented through computer software and result in a technical effect or contribution then the test applicable to computer software can also be applied and patentability can be adjudged.” In other words, the Court will assess the invention for technical contribution only after it is implemented through a computer program. Until then, it cannot be patented.
The Court, in the present case, has taken a leaf out of the Lava Intl. case decided earlier by the DHC this year. There, the Court held that if an algorithm enhances the effect or functionality of a system or hardware, such effect on the hardware will be patentable subject matter. In Para 48, the Court noted that since the invention was purely related to the algorithm, without any substantial change in hardware, it would not be entitled to patent protection.
Blackberry II: Going Beyond Algorithm
Keeping the above legal proposition in mind, it is not hard to reconcile Blackberry I with the Blackberry II case.
In Blackberry II the invention related to a system which organized and managed media content after assessment of a user’s liking. The invention was rejected by the Office, reasoning that the invention was related to algorithms and computer programs per se.
In Para 32, the Court, first, noted that the invention has technical implementation i.e. a combination of hardware and software was used for managing media content. (Para 26) Although the invention contained algorithms or a set of instructions, the algorithms, in combination with a computer program, were implemented using software implementations (Para 27) Therefore, as opposed to Blackberry I, the present invention was not related to purely algorithmic instruction.
Since the algorithm was implemented using a computer program i.e. implemented using software, the test of technical contribution or effect was used. In Para 36, the Court noted that the invention resulted in a significant enhancement of the capability of the device and made it ‘functionally’ more efficient without any changes to storage space (dynamic media synchronization, cache management etc). This, according to the Court, satisfied the test of concrete technical effect and technical contribution.
If one reads Para 52 of Blackberry I (extracted above), it is discernible that the present invention satisfied the conditions laid out in that paragraph. On the whole, the reasoning employed in the above two judgements is consistent.
Patenting Algorithm: Following the Mayo Test
Indian case law on whether algorithm are non-patentable altogether or not is clear: since ‘algorithms’ u/s. 3(k), unlike computer program, is not qualified with the words per se, it is absolutely barred from being patented. A similar reasoning is used in this case.
The question of patentability of algorithm has seen a lot of back and forth in the US. This article discusses Benson, Flook and Diamond– three cases which shaped the discourse on patent-eligibility of algorithm. In Benson, the Court denied protection to a process which relied upon mathematical algorithm due to lack of “substantial practical application.” In Flook, the Court again said that algorithms are natural laws, not man-made product, and thereby unpatentable. However, in Diamond, the Court held that an invention relying upon mathematical algorithm will be patentable if such algorithm is implemented in a structure or process to perform a function which is patentable.
Today, in US, a two-step inquiry is conducted to determine patent-eligibility, called the Alice-Mayo Test–
- Whether the claim of the invention are directed towards a patent-ineligible subject matter, such as mathematical algorithm.
- If yes, does additional elements of the claim amount to significantly more than a claim for the abstract idea i.e. transform the process into an inventive application of the mathematical formula.
In Alice (original judgement),the Court had cautioned that the mere implementation of an abstract idea using a computer will not make it patent eligible. This prevents the circumvention of the abstract idea exclusion under the scheme of patent policy. Rather, there must be “a new and inventive application of an existing formula” to be patent eligible.
The US approach, on one hand, ensures that abstract ideas are excluded from enjoying patent protection, while on the other, grants protection to inventions relying on algorithms as long as it is inventive.
DHC judgement in Lava followed a similar approach, wherein it held that an invention, which merely incorporates an algorithm, is eligible for patent if it results in “enhancing the functionality of a system or hardware component.” The functionality derived, according to the Court, is patentable subject matter. The Algorithm by itself, however, was still unpatentable. (Para 69) Connecting with the Mayo Test, such enhanced functionality “amounts to something significantly more than a claim for the abstract idea.”
Blackberry I, as shown above, uses a similar approach as followed in Lava. Therefore, although an algorithm is ineligible for protection, an invention can still use an algorithm for implementation and seek patent protection as long as it satisfies other criteria for patentability.
A recent post by Bharathwaj argues that Lava, marking a shift from Open TV, has narrowed the exclusion of algorithm patenting, saying it is no longer absolute. As I have highlighted in my comments there, I believe Lava merely clarifies the holding in Open TV, that while the bar against patenting an algorithm is absolute, an invention will not be rendered unpatentable merely for incorporating a business method, algorithm etc. in its implementation. Therefore, in my opinion, there is not a shift in the position and the bar against patenting an algorithm remains absolute.