Behind the Scenes: The Making of the 2017 CRI Guidelines

(2,700 words, approximately ten-minute read.)

In July this year, we covered the Indian Patent Office’s decision to revise its examination guidelines for Computer Related Inventions (“CRI Guidelines”). We noted that it was the third set of guidelines released in as many years – the guidelines were first issued in August 2015, suspended and kept in abeyance until February 2016, when a fresh set was issued which almost entirely contradicted its predecessor.

We have dealt with the substance of the guidelines earlier – the 2015 Guidelines were seen as broadly opening the gates for software claims, while the 2016 Guidelines introduced fresh provisions in Paragraphs 5 and 6 which were perceived to preclude all software claims from patent protection. (Broad generalisations like this are blatantly untrue, but that will be the subject of a future post.) Of particular importance here is the three-step test in Paragraph 5 of the 2016 Guidelines, whose third step required examiners to outright reject the claim if the claimed invention did not possess any novel hardware.

I filed an RTI with the CGPDTM on the day that the 2017 Guidelines were issued to find out more about the Dr. Jekyll/Mr. Hyde behaviour on CRIs, and sought information on the decision-making process involved. This post details the results of that RTI investigation, which raises several questions on the almost whimsical nature of IP policymaking in India today.

Queries

Among the information I requested was the reasons for contemplating a change to the 2016 Guidelines, details of the Expert Committee which recommended the issue of the 2017 Guidelines, the manner in which the EC was constituted, the stakeholder consultations it conducted, the representations made to it, and the report finally prepared by the EC containing its recommendations.

Findings

While my request was directed to the CGPDTM originally, I received a reply on 24 July from his office stating that they did not possess the relevant documents to answer most of my questions, and that my request was being forwarded to the DIPP. I received the DIPP’s reply (dated 10 August) last week, and you can access it here. (66 page PDF, slightly repetitive)

As a kind of TL;DR to this post, you can also find a quick synopsis of the RTI reply (minus analysis) here.

1. Reasons for the rule change

On 18 March 2016, DIPP records that it had received representations from Mishi Choudhary (presumably from SFLC, affiliation unspecified in the document), Jyoti Sagar (K&S Partners), Sheetal Chopra (Ericsson India) and Chander Lall (SSL&S) on the 2016 Guidelines. The DIPP forwarded these representations to the DeitY and the CGPDTM, and proposed that a three-member Expert Committee be constituted to re-examine the 2016 Guidelines. In the file noting, the DIPP also recorded that “Concerns have also been raised in various quarters globally about these [2016] guidelines”. The DIPP proposed that the EC be made up of one nominee each from the DIPP, DeitY and CGPDTM’s office. An Office Memorandum expressing these decisions was prepared on 28 March 2016, and signed by SK Lal of the DIPP. This OM also set 30 April 2016 as the deadline for the EC to submit its recommendations.

Why were stakeholders lobbying with the DIPP instead of the IPO? Is it because the DIPP was perceived to be more applicant/patentee-friendly than the IPO? Where did the DeitY get the mandate to take this decision? What were these “concerns” raised in “various quarters globally”, and why were they persuasive enough for the DIPP to consider their merits? What was SFLC’s role “raising concerns” on the 2016 Guidelines, particularly given that it applauded their publication in the first place? Given that a majority of the representations received seem to have come from a narrow band of “stakeholders” (SSL&S has represented Ericsson in court, as has K&S), how much weight should have been given to their pleas against those of SFLC (which presumably argued in favour of the 2016 Guidelines)?

2. Constitution of the EC

In the same 18 March file noting, the DIPP nominated Sushil Satpute (Director) as its man on the EC. On 1 April, the CGPDTM nominated BP Singh (Dy. Controller) to the EC, and on 8 April the DeitY nominated AK Garg (Director). On 19 April, the EC met and decided that it would base its recommendations on stakeholder consultations.

Is this an instance of the IPO being hustled into removing the novel hardware requirement through a committee in which it would be outvoted by the combined strength of the DIPP and DeitY?

3. Stakeholder representations

On 22 April, the EC conducted a stakeholder meeting. Minutes of these meetings reveal that representations from 19 organisations were considered by the EC. A summary of notable submissions is below:

1. In-house IP Professionals Forum
   • Stakeholders have requested the 2015 Guidelines to be reinstated.
   • 2015 Guidelines are based on the Delhi HC’s interim order in Ericsson v. Intex, while the 2016 Guidelines are not based on any case law
   • Innovation is becoming hardware-agnostic
   • Patents must be granted for software that achieves a technical advancement through non-obvious technical features.
   • Instead of checking whether the claimed invention is implemented purely by software, the test should be whether the claim defines a technical invention that is new and non-obvious.
   • Positive examples are required.
   • The EPO approach of requiring technical effect to be demonstrated for CRIs may be followed.
   • SFLC is wrong to argue that this will result in the grant of patents on software code – patents will only be granted for the underlying ideas or techniques for solving problems, which form the basis of the code.
2. CII
   • Essential to define “technical effect” in the Guidelines.
   • Novel hardware requirement hampers startups because software investments are cheaper than hardware investments.
3. BSA
   • Replace “novel hardware” in Paragraph 5 with “a further technical effect beyond the normal interactions between the programme and the computer”.
   • Lists out software-enabled inventions not involving novel hardware that have already received protection in India
   • Positive examples required.
   • 2015 Guidelines arose out of a deliberative, stakeholder-engaged process. 2016 Guidelines did not.
4. FICCI
   • 3-stage test is inconsistent with Indian and foreign judgements.
   • 2016 Guidelines are contrary to policies in other countries.
5. NASSCOM
   • Paragraph 4.2 should be modified to ensure that even if aspects of the inventive step are statutorily excluded by Section 3, the whole invention must be examined to assess inventive step, in accordance with Section 8.03.03.02d of the Manual.
   • Methods of playing games and computer programming languages must be removed from the section on industrial applicability.
   • Inventions in fields such as encryption, data compression, image processing, etc. may fall outside patent protection simply by virtue of the novel hardware requirement.
   • Software which produces a technical outcome can be distinguished from that which does not, so the novel hardware requirement is overbroad.
   • Apart from the novel hardware requirement, the 3-stage test is in conformity with the Act.
6. SSL&S
   • 2016 Guidelines contravene the legislative history and intent, and were not made after following due process.
   • 2016 Guidelines broaden Section 3(k) to effectively bar all CRIs from patent protection.
   • Positive examples are required.
   • “technical advancement” and “technical effect” need to be defined.
   • 2016 Guidelines jeopardise other initiatives such as Make in India, Startup India and Digital India, and were based on inputs received from organisations with vested interests.
   • 2016 Guidelines violate TRIPS since they prescribe a higher standard for CRIs than other classes of inventions.
7. Pratibha Singh
   • Paragraph 4.4 on sufficiency of disclosure is redundant owing to Section 10 of the Act, which requires that the specification must fully and particularly disclose the invention.
   • 2016 Guidelines are contrary to the Delhi HC’s order in Ericsson v. Intex.
   • Positive examples are required.
8. K&S
   • Positive examples for mathematical methods must be reinstated.
   • 2016 Guidelines are ultra vires the Act and against the Delhi HC’s interpretation of Section 3(k).
9. Anand & Anand
   • No justification given for suspending the 2015 Guidelines.
   • 2016 Guidelines violate Section 2(1)(j).
10. Lakshmikumaran & Sridharan
    • Last two paragraphs of Paragraph 4.4.5 must be deleted, since they suggest that any invention implemented through software is not patentable.
    • Novel hardware requirement must be modified to include novel software coupled with existing hardware.
    • Positive examples required.
    • Inconsistency with Ericsson v. Intex.
11. SFLC (with signatures from ALF, CIS, FSF, FSMI, TWN, Knowledge Commons, Democratic Association for Knowledge Freedom, Internet Society)
    • 2015 Guidelines paved the way for business methods to receive patent protection, and are therefore ultra vires Section 3(k).
    • 2015 Guidelines overlook the unsuccessful Patents Amendment Bill 2005 to extend protection to computer programs with “technical application to industry”. This demonstrates legislative intent against the 2015 Guidelines.
    • 2016 Guidelines are TRIPS-compliant owing to flexibilities in the treaty.
12. iSPIRIT
    • The proposal in 2004-05 to amend Section 3(k) was rejected on the ground that it would give rise to a monopoly of multinationals.
    • Indian Software-Driven Hardware providers do not Indian patents to serve the international market.
13. G Madhusudan (IITM, individual capacity)
    • The market is flooded with inventors claiming algorithms as software.
    • 2016 Guidelines are in line with statutory intent and will prevent the patenting of algorithms.
14. MM Balakrishnan (IITD)
    • Novel hardware requirement is inconsistent with the Act.
    • The hardware/software distinction is blurring with advancements in reconfigurable hardware such as Field Programmable Gate Arrays, which use hardware as a platform and can be programmed to change their function.
    • The 2016 Guidelines may treat like situations in an unlike manner – e.g. Application-Specific ICs would be classified as novel hardware but FPGA prototypes may not, despite the fact that the novelty in both cases resides in the algorithm instructing the hardware.
    • Patents Act must be updated once in a decade to keep pace with technology.
15. Huawei
    • The novel hardware requirement severely hampers innovation in the ICT sector.
    • 2015 Guidelines were well-received by industry, so the re-issue of Guidelines in 2016 is inexplicable.
    • The 3-stage test is unprecedented in Indian jurisprudence and therefore difficult to implement.

Apart from these, Ericsson, Wipro, Majumder & Co., and the US-India Business Council all argued that the novel hardware requirement ought to be removed.

In addition, the DeitY held its own stakeholder consultation on 4 May 2016. Most stakeholders reiterated their earlier arguments, but one additional comment is recorded in the minutes. Anand and Anand seems to have made an interesting argument: that the IPO has no power to give meaning to undefined terms in the Patents Act through guidelines. A&A submitted that just as it was necessary for the Supreme Court to define “efficacy” as “therapeutic efficacy” in Novartis v. Union of India rather than the IPO, the patent office could not conduct statutory interpretation by issuing examination guidelines.

Why did the DeitY conduct its own, separate stakeholder meeting? Again, questions surrounding the DeitY’s mandate crop up. In addition, the Anand & Anand’s submission to the DeitY seems to be an entirely reasonable proposition – while it’s true that the patent office necessarily must take policy decisions at some level (i.e. when granting/refusing individual applications), this realist perspective does not condone the IPO going out of its way to fill in the gaps in Section 3(k) through its examination guidelines (which constitute institutional policy-making). This should ideally be left to be settled through litigation by individual applicants and their opponents.

4. DeitY recommendations

Following its stakeholder consultation, AK Garg of the DeitY sent its recommendations to Sushil Satpute of the DIPP on 27 June 2016. Interestingly, this OM is not copied to BP Singh of the CGPDTM’s office. The DeitY makes three recommendations:

1. Software in conjunction with hardware resulting in novel configuration can be patentable subject matter.
2. DeitY recommends the removal of novel hardware requirement because it “supports innovation-driven ICT industry backed by IPRs” and because the ICT industry is moving towards software-induced performance improvement.
3. The distinction between software and hardware has blurred with reconfigurable hardware such as Field Programmable Gate Arrays.

Either this communication (and the recommendations made by the DeitY, and by extension the stakeholder consultation that resulted in them) falls entirely outside the EC’s mandate, or this communication was made in the course of the EC’s mandate but only from one member to another. In any case, the CGPDTM’s exclusion from this exchange is puzzling, to say the least. In addition, the DeitY’s blanket support for “innovation-driven” industry “backed by IPRs” provides further fuel for my conspiracy theory that its involvement in the process was driven by patentees who wanted a sympathetic ear in government.

5. EC recommendations

The EC submitted its initial report on 30 June 2016, with ten recommendations. Below are the really spicy ones:

1. Rephrasing and rearrangement of the inventive step commentary in Paragraph 4.2 in line with NASSCOM’s suggestion.
2. Paragraphs 4.4.1 and 4.4.4 to be modified to remove references to computer memory and processors, to avoid confusion.
3. Insertion of positive examples for mathematical methods, in accordance with K&S’s suggestion.
4. Modification of Paragraph 4.5.4 to specifically include claims whose contribution lies in “a novel configuration resulting from a combination of software and hardware and leading to enhanced functionality” as patentable subject matter.
5. Removal of Paragraph 5 (3-stage test, including novel hardware requirement) altogether since they are “mere repetition of the issues well-deliberated under novelty, inventive step and other specific chapters in the Guidelines”.
6. Inclusion of positive examples.

6. DoT cameo

Over four months later, on 17 October, Ramesh Abhishek, Rajiv Aggarwal (Secretaries, DIPP), Aruna Sundararajan (Secretary, MeitY) and JS Deepak (Secretary DoT) meet to review the EC and DeitY recommendations. They conclude that:

1. Section 3(k) does not prescribe novel hardware.
2. “novel configuration” as recommended by the EC must be defined in the new Guidelines
3. EC’s recommendation to exclude indicators to determine technical advancement (2015 Guidelines) must be accepted.
4. Positive examples must be included.
5. EC report to be sent to CGPDTM for action.

Why was the Department of Telecom involved in making the rules governing software patenting?

7. Amendments and further stakeholder meetings

Nearly four more months elapse, and on 14 February 2017 the DIPP requests the EC to re-examine the terminology used in recommendation (vii) [(4) above]. On the very same day, the EC replies, amending the language to cover claims whose technical contribution “lies in the combination of software and hardware that results in the technical advancement in the configuration and leads to enhanced functionality”, with “configuration” being defined as in the Oxford Advanced Learner’s Dictionary.

Another four months later, on 7 June 2017, the DIPP informs the EC of stakeholder consultations that it has conducted based on the EC’s report. DIPP says that there is a “diversity of opinions” on two points. First, some stakeholders (presumably from the open-source community) argued that recommendation (viii) didn’t account for the failed attempt to amend Section 3(k) in 2004. Second, some stakeholders resisted recommendation (x) [(6) above] which required positive examples to be inserted, illustrating patentable subject matter.

The EC replied on 19 June 2017, and conducts a complete volte-face on recommendation (x), asking for examples to be omitted altogether. On recommendation (viii), it refuses to budge.

Epilogue

On 30 June 2017, the revised CRI Guidelines were published. In pure numerical terms, 16 of 19 submissions argued against the 2016 Guidelines before the EC, and 11 of 24 stakeholders did so before the DeitY. We do not know if the IPO was strong-armed into this decision through strategic lobbying with sympathetic government organs, or whether the 2017 Guidelines reflect a genuine change of heart on the patent office’s part. In any case, the 2017 Guidelines do not, by any means, settle the position of Indian law on software patents. Neither the Guidelines themselves (or their predecessors, for that matter), nor the Office Order that promulgated them provide any substantive reason for the rule change – they merely enact the farce that the Guidelines are only clarificatory and that the law has always been static.

Unfortunately, whether or not the Guidelines possess any legal value, the fact is that they will be relied upon by applicants, their agents, and (more importantly) by examiners to varying extents. Inasmuch as they form institutional policy that is not backed by substantive reasoning, they carry about as much merit as the NALSAR Hostel Rules did in 2013 when I entered University, which banned the possession of alcohol, tobacco, narcotics and camera phones. It is 2017, and the NALSAR Hostel Rules are currently being revised by a body competent to formulate authoritative, binding and consistent policy (and is comprised of students, to boot!). This can only happen with Section 3(k) through litigation or parliamentary intervention, and I doubt either is going to happen anytime soon.