October 31, 2018

Facts

The claimant, Clearswift Ltd, which was an anti-malware provider, challenged the validity of a patent owned by the defendant, Glasswall (IP) ltd.  The patent was the UK designation of a European Patent entitled, “resisting the spread of unwanted code and data.”

Clearswift sought revocation of the patent under s 72(1)(a) of the Patents Act 1977 on the basis that the patent lacked inventive step over two items of prior art.

David Stone, sitting as a Deputy Judge of the Chancery Division, gave some background to the case, explaining that the rapid increase in internet use (and particularly email use) in the 1990s had led to an increase in the proliferation of malware, including computer viruses, and a concomitant proliferation of anti-malware software to try to prevent it.

As at the priority date of Glasswall’s patent in 2005, the majority of anti-malware software worked by scanning for malware, or evidence of the presence of malware.  Incoming files (including emails) were scanned, and blocked (or parts of them blocked) if evidence of malware was detected.  Scanning systems could only look for what they knew, meaning that new malware could not be detected until the anti-malware software had been updated.

Glasswall’s patent claimed to take “an entirely different approach to protection against unwanted code”.  Rather than looking for known “bad” code, Glasswall said that the patent taught “parsing” (breaking down data into its component parts) and “regeneration” of each file, in effect breaking it down, extracting its content and rebuilding it according to known rules.  This meant that only “good” code was passed on to the recipient.  Any “bad” code was left behind, but the regeneration occurred regardless of whether any malware was identified.  The patent also included a threat filter.

In response, Clearswift cited a US patent published seven weeks before Glasswall’s patent, and a series of posts on an internet bulletin board in which various experts discussed the Glasswall anti-malware software that was available at that time.  Clearswift said that there was no inventive step from the US patent or the expert internet discussion to Glasswall’s patent.  It said that the US patent and Glasswall’s patent were doing the same thing in the same way.  As for the expert internet discussion, Clearswift said that it disclosed to the skilled person the whole of the relevant claim of Glasswall’s patent.

There were no witnesses of fact.  Each party called a single expert witness.

Expert witnesses

Clearswift’s expert gave well-considered evidence as to the common general knowledge of the skilled person, as at the relevant time.  However, his evidence was criticised by Glasswall, who said that he could not reliably distinguish between what was known, and what was common general knowledge.  Nor could he reliably distinguish between what could be done and what was obvious to do.  Glasswall said that this “coloured” his evidence.

Clearswift criticised Glasswall’s expert because he had never worked in industry, and had not in his academic career worked on or supervised anti-malware projects.  Glasswall conceded that Clearswift’s expert was better placed to assist the court, but said that its expert had been able to educate himself from contemporaneous materials so as to be able to help the court.

Mr Stone said that there was some force in the criticisms by each party, but that they only went so far.  He followed Lord Justice Jacob’s approach in Technip France SA’s Patent [2004] RPC 46, in which he had said that the primary function of an expert witness in a patent case was “to educate the court in the technology – they come as teachers, as makers of the mantle for the court to don.”  Therefore, Jacob LJ had said, it did not matter whether they approximated to the skilled person.  “What matters is how good they are at explaining things”.

In Mr Stone’s opinion, both experts in this case were good at “explaining things”.  Both sought to help the court and both were able to fulfil the role of teacher.  Neither was (or could be) an approximation for the skilled person: one because he was significantly too skilled and inventive, and the other because he had never worked in industry.  However, as Jacob LJ had said, that was not their role.  Mr Stone was satisfied that both experts were able to assist the court in understanding the technology.

Mr Stone also said that where the experts disagreed, it was not a matter of simply accepting one over the other because he was a better witness.  All the relevant determinations (the construction of the patent, the teachings of the prior art, and whether the patent involved an inventive step) were for the court to make, based on the evidence.

Decision

In terms of assessing obviousness, Mr Stone followed Pozzoli SPA v BDMO SA [2007] FSR 37, in which Jacob LJ said:

1. i) identify the notional “person skilled in the art”;
2. ii) identify the relevant common general knowledge of that person;

• iii) identify the inventive concept of the claim in question or if that cannot readily be done, construe it;

1. iv) identify what, if any, differences exist between the matter cited as forming part of “the state of the art” and the inventive concept of the claim or the claim as construed; and
2. v) viewed without any knowledge of the alleged invention as claimed, do those differences constitute steps that would have been obvious to the person skilled in the art or do they require any degree of invention?

As for the skilled person, Mr Stone said that this was a person with a computer science degree and between one and three years’ experience working in the anti-malware software industry.

The relevant common general knowledge of the skilled person was, essentially, the anti-malware techniques in use at the relevant time as a result of the increase in the proliferation of malware, including computer viruses.

As for establishing the inventive concept, Mr Stone again followed Pozzoli, which said that, where there was disagreement on the issue, the court should work on the features of the claim.  Construing the patent, Mr Stone preferred Glasswall’s interpretation.

In terms of the prior art, Mr Stone found that parsing and regeneration were key elements of Glasswall’s patent, and they were not found in the US patent.  This was “a significant difference that would not have been obvious to the skilled addressee.”  Further, the threat filter in Glasswall’s patent operated in a different way to the US patent and was therefore inventive.

As for the internet bulletin board discussion, Mr Stone agreed with Glasswall that it was difficult for the skilled person to extract much from it that was “clear and unmistakable.”  It did not disclose a number of the teachings of Glasswall’s patent.

Accordingly, Mr Stone found that the patent was not obvious over the prior art and Clearswift’s application for revocation of the patent was dismissed.  (Clearswift Ltd v Glasswall (IP) Ltd [2018] EWHC 2442 (Pat) (28 September 2018) — to read the judgment in full, click here).