Third time's a harm? Microsoft tries to get twice-rejected compression patent past skeptical examiners

Boffin who invented ANS encoding and made it public domain worries Microsoft will come rent-seeking


Analysis In June, 2019, Microsoft applied for a US patent covering enhancements to a data encoding method known as rANS, one of several variants in the Asymmetric Numeral System (ANS) family that form the foundation of data compression schemes used by Apple, Facebook, Google, various other companies, and open source projects.

Its US patent application was published on the last day of 2020. Recently, the inventor of ANS, Jarosław Duda, assistant professor at Institute of Computer Science at Jagiellonian University in Poland, expressed concern that if Microsoft's patent application is granted, anyone using software that incorporates an ANS-based encoder could be at risk of a potential infringement claim.

"If granted, it will greatly damage ANS applicability and development," Duda explained to The Register in an email.

The Register asked Microsoft to elaborate on why it is seeking an rANS-related patent,and the company declined to comment.

This is not the first time a large tech company has sought an ANS-related patent. Google tried to get a patent for an ANS-based video compression that it ultimately gave up on after reports about the company's effort to claim ownership of public domain work and correspondence from Duda led the US Patent and Trademark Office (USPTO) to reject the initial application in 2018.

patent

'Tech troll' sues EFF to silence 'Stupid Patent of the Month' blog. Now the EFF sues back

READ MORE

"Google ended up abandoning that application," said Alex Moss, staff attorney for the EFF and Mark Cuban Chair to Eliminate Stupid Patents, in an email to The Register. "But it looks like Microsoft picked up right where it left off."

"Professor Duda’s concerns about the Microsoft application are similarly well-founded: these are broad claims that implicate practically any use of ANS without adding anything new and non-obvious," said Moss.

The USPTO has already said as much, Moss explained: It has rejected this application twice before, including a final rejection for obviousness.

The USPTO issued a non-final rejection of the application on May 21, 2020. Microsoft sought a review of the decision and the patent agency then issued a final rejection on October 27, 2020.

Yet on March 2, 2021, Microsoft tried one more time to get its patent application approved. In a USPTO explanatory filing, attorney Kyle Rinehart said, "The Applicant respectfully disagrees with the rejections."

"Microsoft’s recent filing takes advantage of what’s called the "After Final Consideration Pilot 2.0" program," Moss explained. "This program was started under former Director of the Patent Office, Andrei Iancu, and before leaving office, he extended the program through September 30, 2021."

Iancu, said Moss, "went out of its way to tilt the scales in favor of patent applicants whenever possible," and pointed to criticism of the USPTO she penned last year.

XL claims coming with JPEG

The rANS variant, Duda said, is used in JPEG XL, an image compression format, backed by Google and others, that's rapidly gaining support because it provides three times better compression than the JPEG specification developed in 1992.

"This patent will make JPEG XL adaptation more difficult, so I hope [Google's] team will help fight it," he said. "But generally this situation shows again that people giving away their work can often become victims of the patent systems."

Duda argues that just keeping track of all the attempts to patent variations on existing work makes it a challenge to defend the public domain against "patent vultures."

As an example, he pointed to the CRAM file format, a compression scheme based on rANS that's used for storing DNA sequence data. The code authors at the Wellcome Sanger Institute in the UK made CRAM available for free and "now have a full time job of defending it from patent vultures trying to get monopoly on its already used methods," lamented Duda.

Beyond the USPTO's prior rejections of Microsoft's patent application, there's skepticism among developers who work with ANS that what Microsoft describes is really innovative enough to qualify for a patent. An online discussion of the patent last week involving Duda, James Bonfield (a software developer at the Wellcome Sanger Institute and author of CRAM's C implementation), Jon Sneyers (senior image researcher at Cloudinary and editor of the JPEG XL spec), and others casts doubt upon whether anything described in the patent would qualify for protection.

But many unworthy patents applications get granted and getting them re-examined and revoked tends to be costly and often involves litigation.

Patent battle continues

Bonfield told The Register that Microsoft's patent would clearly pose a problem if granted. "Trying to prove a patent was wrongly granted is a minefield and not something most individuals could do, only large corporations, so they can be a big problem for hobbyists and academia," he said.

He added that he believes software patents are wrong. "They're disallowed in the European Union, but in a global world if one country permits them then it can still cause damage, especially when they involve data interchange."

Sneyers told The Register via email that he didn't have any reason to believe Microsoft's patent application would affect JPEG XL since there hasn't been an ISO patent declaration by Microsoft. The only two declarations on JPEG XL, he said, have been from Google and Cloudinary, both of which have said they have related patents and have made them available on a royalty-free basis.

He added however that he's aware Duda wants his ANS invention not to be patent encumbered and that he considers that a noble stance. "...it would be a sign of intellectual integrity [on Microsoft's part] to not attempt to sabotage that by patenting what appear at a first glance to be straightforward and not particularly innovative variations on his invention," he said.

"I personally am of the opinion that software patents, at least the way the current system works, do not have a particularly beneficial overall impact on innovation. It is more of an obstacle than an incentive to innovation," said Sneyers.

"This is made worse by the relative ease of getting bogus patents granted, the legal costs and barriers to get bogus patents invalidated, and the common practice by patent trolls to threaten smaller companies with litigation and getting settlements even when the patents are clearly bogus or do not apply, just because settling is cheaper than fighting them."

Moss said concerns about wrongly-granted patents are a matter of genuine concern.

"That’s especially true when they cover and therefore would restrict access to ubiquitous technologies that are necessary for interoperability, like ANS," she said.

"The Patent Office is supposed to promote innovation and access to knowledge, not the proliferation of patents. Where, as here, a patent application covers technology that is well-known and widely-used, granting it only hurts the public by depriving them of access to knowledge they already had."

"We share Professor Duda’s concerns about this application, and hope the departure of former Director Iancu will give the Patent Office another chance to examine it carefully and reject it once again," she said. ®


Biting the hand that feeds IT © 1998–2021