Thursday 31 August 2023

Bitcoin File Format

Does copyright subsist in the 'Bitcoin File Format'? This was the question in Wright & Ors v BTC Core & Ors [2023] EWHC 222 (Ch), where Mellor J found earlier this year that it did not because the subject-matter was "not expressed or fixed anywhere" (paragraph 62). On appeal ([2023] EWCA Civ 868), Arnold LJ decided that all that was required was that "the structure be completely and unambiguously recorded" (para 69). The claimants did therefore have a "real prospect of successfully establishing the fixation requirement is satisfied" and the appeal was allowed. Do the claimants now therefore have a prospect, real or otherwise, of establishing that this fixation requirement is satisfied?

Firstly, for those who may be unfamiliar with the term (which is possibly understandable since it appears to have been invented for the legal proceedings), what is the "Bitcoin File Format"? According to the claimants, it is "the original work consisting of the structure of each block of the Bitcoin Blockchain" (referring to a Schedule which is unfortunately not publicly available, but which was written in 2022). The Bitcoin File Format (BFF) is therefore the way that data is arranged in each Bitcoin block. The basics of this are explained on the Bitcoin Wiki page for the Genesis Block, which in raw hexadecimal form looks like this:

Genesis Block (annotated)

My annotations indicate the various parts of the block, which have a particular meaning when read by the Bitcoin software. There is therefore a clear format to the block, which is the same for every subsequent block, but importantly the format is not clear from the block itself. There is in fact no information in the block itself about what each section means. Instead, the meaning has to be inferred from software that knows what to look for. This makes the fixation of the format different to that in, for example, an XML template where the template contains a structure (in <brackets>) that indicates what the content is about. This allows an XML reader to extract information from a file created using that template and order it accordingly. Previous cases have found that copyright can subsist in XML templates, i.e. an XML file with just the placeholders and no content.
An example XML template (from here)

This was the difficulty that Mellor J had when assessing whether the claimants had established that there was a serious issue to be tried. If the claimants could not establish how the format itself (not simply the data output, which was inevitably different every time) was fixed, the fixation issue was not met and there was no serious issue to be tried. The judge attempted to get the claimants to file evidence showing how the fixation requirement was met. The claimants filed various explanations from themselves and third parties that showed how the structure of each block was defined, but none of this was found to be relevant because they were not relevant works identifying the structure, only things that were provided after the event. As Mellor J stated:
"It is most revealing that, despite all these opportunities, the Claimants have not filed any evidence to the effect that a block contains content indicating the structure, as opposed to simply reflecting it. By 'content indicating the structure', I mean, by way of a crude example, a flag or symbol in the block which signals 'this is the start of the header' or 'this is the end of the header', or an equivalent of the sort of content which is found in an XML file format. Whilst I entirely accept that each block conforms to the structure described in Schedule 2 to the Particulars of Claim and is an instance or manifestation of that structure, the absence of such evidence confirms my initial view that, whether one considers the point at which the first, second or subsequent block(s) were written embodying the structure of the file format, nowhere was the structure of Bitcoin File Format fixed in a copyright sense in a material form in any of those blocks" (paragraph 57).

This is the key point on which the judgment was overturned on appeal. It is indeed clear that the format of the block is not evident from the block itself, unless you know what you are looking for, but Arnold LJ considered that this did not actually matter. The block did have a format, but the fixation of this format did not need to be in the same place. He summarised the requirements that the claimant would need to meet in order to establish that copyright subsisted in the BFF, which were set out (in paragraph 61) that: 

i) the Bitcoin File Format is a work; 
ii) it is a work that falls within one of the categories of protectable work specified in the 1988 Act; 
iii) the work has been fixed; 
iv) the work is original; and 
v) the work qualifies for copyright protection under the 1988 Act.

The only point in contention was iii), which Mellor J found was not met. Arnold LJ considered that, while it was correct that the work, that is to say the structure, must be fixed in order for copyright to subsist in it, it did not necessarily follow that content defining the structure was required in order to fix it. All that is required is that the structure be completely and unambiguously recorded. The claimants did therefore have a real prospect of successfully establishing that the fixation requirement was satisfied, provided they were able to provide evidence of the structure being recorded from the right time, i.e. from before the alleged infringements. 

An odd feature of this case is that, in the original Particulars of Claim (PoC), the claimants alleged infringement of: i) Database Right in the Bitcoin Blockchain (discussed by me here); ii) Copyright in the Bitcoin File Format; and iii) Copyright in the Bitcoin White Paper. There was no mention of the Bitcoin software itself, which as it turns out is actually where the Bitcoin File Format is fixed. The structure of the Genesis Block, for example, is in fact completely defined in the original Bitcoin software. It therefore strikes me as strange that the claimants chose not to refer to this as providing the fixation requirement, particularly as Mr Wright claims to have written the software himself and would therefore presumably know where to look. Given that Arnold LJ has decided that the work and the fixation requirements can be met by different things, perhaps this will be what the claimants come up with next in their attempts to enforce copyright in the Bitcoin File Format. 


Friday 25 August 2023

Heraeus Noblelight v First Light Lamps

Since the case of Catnic Components v Hill & Smith came before the House of Lords in 1978, the principle of "purposive construction" has been established in UK law regarding interpretation of patent claims. In the Catnic case, the issue was about whether a patent claim to a lintel having a support member "extending vertically" covered an otherwise identical lintel having a support member that was not quite vertical. According to Lord Diplock's leading judgment, "A patent specification should be given a purposive construction rather than a purely literal one derived from applying to it the kind of meticulous verbal analysis in which lawyers are too often tempted by their training to indulge". This meant that the term "extending vertically" was to be purposively construed to effectively mean that the member should be sufficiently vertical for it to serve its purpose. A few degrees either way would still allow it to serve its purpose, so the term in claim 1 of the patent was not to be construed narrowly and the defendant's lintel was found to infringe. 

Even though the law has since developed into what is now, following Actavis v Eli Lilly, effectively a doctrine of equivalents for patent claims, the principle of purposive construction has not gone away and is still in use when it comes to interpreting specific terms in a patent claim in a way that means they should perhaps not be construed as narrowly as they first appear. 

A nice example, again with a simple mechanical invention, has recently appeared in the form of the judgment of Mr Justice Zacaroli in Heraeus Noblelight Limited v First Light Lamps Limited [2023] EWHC 1950 (Pat), issued on 31 July 2023. The case related to Heraeus' European patent EP1598845B1, which claimed a method for forming a quartz glass lamp, the invention relating to how a seal between a metal electrode and a quartz glass tube was formed during manufacture of the lamp. The problem the invention aimed to address was the previously complicated method of manufacture that involved many different stages and sometimes still resulted in failure of the seal between the electrode and the quartz glass. 

It was already known that an intermediate glass material having a thermal expansion coefficient between that of the electrode and the quartz could be used to reduce the differential stress set up on cooling the seal, but the manufacturing process was still complicated. The invention simplified this by, instead of making two seals with the intermediate glass that were subsequently joined together, forming a direct seal using a bead formed on the electrode that joined directly to the quartz tube, following which an internal pressure in the tube caused a smoothing out of any edges internally. This is best illustrated with two simple drawings from the patent, reproduced below. In the first drawing (Figure 2 of the patent), the electrode 2 has a bead 4 of intermediate glass material around its electrical feedthrough 3. This is inserted into a quartz tube 1 while heat is applied and the electrode is rotated relative to the quartz tube. After the glass bead is sealed against the quartz tube, as shown in the second drawing (Figure 1 of the patent), an internal pressure is applied to the quartz tube to smooth out the internal join with the glass bead, forming a smooth internal radius that reduces the possibility of cracking on cooling. 

Electrode (2,3) with bead (4) prior to sealing with quartz tube (1)

Electrode after being sealed against quartz tube.

The patent claimed the method in the following way (with emphasis added on the key feature in question):

1. A method of forming a quartz glass laser lamp or quartz glass flash lamp containing electrodes (2) at opposite ends within a tube (1) of quartz glass, an end of said tube (1) having an internal diameter, an external diameter and an annulus, wherein said method comprises formation of a glass to metal seal at said one end of the quartz glass tube comprising the following steps:
    (1) the electrode (2) or its electrical feed through (3) is surrounded with a sealing glass (4) comprising a sheath and a bead, the bead being bigger than the internal diameter of the end of the tube (1) but not bigger than the external diameter of the end of the tube (1),
    (2) heating the bead to a soft state while rotating it,
    (3) inserting the bead into the annulus so as to form a seal,
    (4) after the insertion, heating the seal so that the sealing glass (4) wets on and fuses with said one end of the tube (1), both internally and to the end of the tube (1),
    (5) after said fusing of the sealing glass (4) with said quartz tube (1), applying an internal positive pressure while the sealing glass (4) is molten causing the sealing glass (4) inside the tube (1) to move back towards the end of the tube (1) so as to form an internal radius of sealing glass (4) which is directly sealed between the electrode or its electric feed through (3) and the tube (1) of quartz glass.

It is clear that it would be necessary for the bead to be bigger than the internal diameter of the end of the quartz tube, otherwise a seal would not be made. However, did the claim really require that the bead was literally "not bigger than the external diameter of the end of the tube"? The question was important because the alleged infringing method did involve making the bead bigger, only by around 0.1-0.3 mm, but otherwise involved forming a seal in the same way as claimed in the patent.  

Heraeus argued that the term "not bigger" should be interpreted purposively as "not bigger by reference to tolerances that formed part of the common general knowledge [CGK]", claiming that 0.3 mm was within such tolerances. Alternatively, the term should be interpreted to be "not bigger such that it impairs the strength or utility of the seal", encompassing oversized beads. They also argued for literal infringement on the basis that some of the beads used by the defendant would inevitably be not bigger than the quartz tube due to variations, or alternatively that according to the doctrine of equivalents a lamp manufactured with oversized beads would be an equivalent. 

Having heard the arguments from both sides, including expert evidence, the judge found that the skilled person's CGK would include beads of varying sizes, including oversized beads. The question was then whether, as a matter of construction, the words "not bigger than" the outer diameter of the tube was to be read as imposing a stricter tolerance than that which the skilled person would understand to be CGK. The claimant argued that it did not, while the defendant argued it did. The judge considered that, since the requirement for the outer diameter of the finished lamp did not generally require such a tight tolerance, the protection sought in the patent was not limited to a method in which beads were precisely no bigger than the outer diameter of the tube but included variants in which the bead was not materially bigger than the outside diameter of the tube to an extent "which could have no material effect upon the way in which the invention worked" (per Lord Diplock in Catnic). Bigger in this case therefore meant not bigger such that it impaired the strength or utility of the seal, which encompassed any bead made within CGK tolerances. 

For completeness, the question of infringement by equivalents resulted in the same conclusion. Although it was not necessary to consider the question of literal infringement, this was determined not to be found because the evidence did not show smaller beads actually having been used. 

The patent was found to be valid over the prior art cited by the defendant (the arguments for which I will skip over here as they don't seem to me to raise anything of interest), so the patent was found to be infringed by the defendant's process on both a purposive construction of the claimed invention and according to the doctrine of equivalents. 

You might think that the lesson to be learned from this would be that an overly specific patent claim can still be infringed even if some features are not literally present in the alleged infringement, and you would be correct. However, the trouble and expense caused by including such a specific feature in a patent claim can nevertheless be enormous if such a claim is litigated. Patent attorneys, being generally conservative and not wishing to cause unnecessary trouble, will most likely take a different lesson, which is that specific terms in a claim that are not strictly necessary for the invention to work or to provide a distinction over the prior art should be avoided. Following Catnic, drafting practice did not change so that specific terms such as "vertical" were used by patent attorneys, happy in the knowledge that they could be interpreted to mean "vertical enough to do the job required". Instead, the word "substantially" was used ubiquitously to provide more wriggle room, often without realising the reason behind doing it. In a similar way with this case, I would guess that any patent attorney reading it would have a more careful think about the claim in front of them and perhaps try not to use any limiting terms that really do not need to be there. I certainly will.

Wednesday 23 August 2023

Satoshi or Not - Does it Matter?

Apparently as part of the build-up to the COPA v Wright hearing at the High Court early next year, an article appeared last month in Forbes (paywalled unfortunately, but easy to work around if you know how HTML works) with the title "Satoshi or Not, Here He Comes". The article is largely about Craig Wright, who is described rather charitably by the writer as an "Australian computer scientist". As an aside, if that definition applies then perhaps I should describe myself as a British computer scientist, based on my extensive experience of programming in BASIC on a ZX Spectrum in the 1980s. 

The writer of the article interviewed Mr Wright himself as well as a few others including me (in my human persona). The main thrust of the article is that, as the title suggests, it may not matter whether Wright is proven to be the person behind the pseudonymous inventor of Bitcoin, Satoshi Nakamoto. This is because the company he has been working for over the past several years now has an enormous patent portfolio that could soon be a threat to many. As the writer states: "if he can wield his trove of 800 granted and 3,000 pending patents in 46 jurisdictions the way he wants to, he could soon start charging for the right to build a wide variety of blockchain applications. That would affect everything from the $1 trillion cryptocurrency market to corporate implementations built by some of the largest companies in the world". That is quite a big "if", which seems to assume that Wright will still be around to wield anything after losing his court case as well as there being some value in the patent portfolio if he is. So, let's imagine what might happen if (as seems practically certain to anyone who is not being paid to say otherwise) Wright fails to prove to the High Court next year that he is Satoshi. This would primarily mean that it will become clear to everybody, if they were not previously aware, that Wright does not have, nor has he ever had, any special knowledge about Bitcoin. The patent portfolio should not therefore have any special status anymore, if it ever did have. Another consequence of this would be that, if Wright has been advancing a deliberately false case (which he has been known to do), this would cast serious doubts on the quality and provenance of any claimed inventions in the portfolio. Another point worth getting into is that Wright's so-called "Satoshi Vision" version of Bitcoin, BSV, is the basis of pretty much everything his company nChain has been working to build on since 2016. Most of the patent applications that have been filed since 2016, which now exceed 500 (based only on the number of GB priority filings) are based on the principles on which BSV is based, which are in important ways fundamentally different to those of the original Bitcoin. A substantial part of the patent portfolio's value would therefore be dependent on the principles of operation of the BSV network itself being valuable and something that others would want to use (for example for CBDCs, cattle tracking or  duplicate pictures of cute dogs, among other things). Unfortunately for BSV, this is evidently currently not the case, at least based on the current market value of BSV.

Going back to the article, an impression that it may have given (which I am sure was intended) is that, since nChain are being granted patents, the patents must be valuable. I was even contacted by someone very senior in the Bitcoin space who asked me directly why I seemed to be saying that the patents had value when, in his view, the portfolio was virtually valueless. I was indeed accurately quoted in the article as saying "A lot of these patents are, for better or worse, valid". This did not of course mean that I thought they were valuable. My view is, in fact, quite the opposite based on my own research on the European part of the portfolio, which at the time of writing amounts to 77 granted patents and 376 pending applications. To anyone familiar with how the patent system works, with a sufficiently technical-sounding specification and enough time and money, almost anything can be forced through the patent system so that it turns into a granted patent. This does not mean that the resulting patent has any value. If the claims are amended so that they cover a narrow technical implementation to work around prior art cited by a competent examiner, even if they cover what the applicant actually does in practice, the resulting patent is inevitably easier to work around and infringement can be avoided, making licensing tricky. For the few applications that do cover something broader and somehow work their way past an examiner (which has happened in a few cases), the resulting patents are more likely to be vulnerable to attacks in post-grant proceedings. What is left is effectively a lottery where an invention might be hidden amongst the trove that, by chance, turns out to be useful and valuable, but this would need a lot of persistence, focus and money to find. Given the enormous cost of prosecuting and maintaining a portfolio of hundreds of applications and patents in just one jurisdiction, it would need to take a world-beating invention worth at least hundreds of millions to justify the expense. It might, of course, be in there somewhere but I wouldn't bet on it.

Sunday 13 August 2023

Third Party Observations

Third party observations can sometimes be a useful tool when keeping an eye on other people's pending European patent applications. While EPO examiners typically do a decent job of search and examination, there are exceptions and even the best examiners only have a limited amount of time to deal with applications so cannot always cover all possibilities. Sometimes a little prod from a third party can make the difference, for example the difference between a third party having to file an opposition to get rid of a troublesome patent and preventing the patent from being granted in the first place (or at least being granted with a limited and non-threatening scope). 

Knowing that examiners typically do their job well, when is the best time to consider whether third party observations should be filed? If you have knowledge of highly relevant prior art, the best time is probably while examination is underway, preferably after the search report issues and before substantive examination begins. Observations are then more likely to be taken into account and the applicant can be forced to narrow the claims to work around the prior art. This is also relevant for other issues such as added matter and clarity, although the examiner should really be on top of these so should ideally be left to do their job. What, however, if there are issues that don't become relevant until later on? Things can be a bit more difficult, which I will show using a couple of recent examples. 

The relevant section of the Guidelines is E-IV, section 3, which states:

Following publication of the European patent application under Art. 93, any person may present observations concerning the patentability of the invention. Although lack of novelty and/or inventive step are the most common observations, third-party observations may also be directed to clarity (Art. 84), sufficiency of disclosure (Art. 83), patentability (Art. 52(2) and (3) or ) and unallowable amendments (Art. 76(1)123(2) and 123(3)).

So there is no restriction on who can file observations as well as no restrictions, unlike in opposition proceedings, on filing observations regarding clarity. Another issue that is not specifically mentioned is that there is no restriction on observations regarding double patenting (G-IV, 5.4), which is also not a ground of opposition. 

In theory, observations should be taken into account at any stage while an application is still pending, i.e. before a decision has been made to refuse or grant the application. In practice, however, there is an issue with observations that are filed after a communication of intention to grant under Rule 71(3) EPC has issued. According to the Guidelines:

Observations by third parties received in examination after dispatch of a Rule 71(3) communication but before the decision to grant (EPO Form 2006A) has been handed over to the EPO postal service will be considered by the examining division. If they are relevant, the examining division will resume examination. Otherwise, brief substantive feedback will be provided in the file.

There should, therefore, be no difference between observations filed before or after issuance of such a communication if they are relevant. In practice, however, observations filed after a Rule 71(3) communication are typically far less likely to result in further examination largely because, at that stage, an examiner has already got to the point where, as far as they are concerned, examination is finished and there is apparently little to no incentive to reopen examination. There may, in fact, be an incentive not to reopen examination if the examiner will have to spend more time on the application without getting any more points. Any observations would therefore have to raise a pretty serious issue for examining proceedings to be reopened. 

As a first example, let's say a divisional application is examined and goes straight through to allowance. Claim 1 as allowed in the divisional is shown below, along with claim 1 of parent application as allowed (and since granted). Notice any similarities? Clearly in this case the examiner had not examined compliance with the Guidelines on double patenting, particularly following G 4/19. This should not have happened, and third party observations clearly could not have been filed before issuance of the notice of allowance to bring this to the examiner's attention. Whether the applicant or their representative would notice and point this out to the examiner is impossible to tell, but they would presumably not have an incentive to do so, particularly if their main incentive is to get as many granted patents for their client as possible without much regard to quality or cost.  

Claim 1 of allowed divisional.

Claim 1 of granted parent.

So, what should be the effect of a third party filing observations pointing out the double patenting requirements to the examiner in this case? The Guidelines state that examination should be reopened, but only if the observations as "relevant" and, if they are not relevant, the examiner must provide a note to this effect on the file. In this example, it is clearly impossible to argue with a straight face that double patenting is not a relevant observation, so the examiner must reopen examination, which it turns out is what the examiner did. The communication of intention to grant was rescinded and a communication under Article 94(3) EPC issued pointing out the requirement to the applicant and asking them to make amendments. Examination for that application is consequently still underway as a direct result of third party observations being filed, and the applicant has indeed made some narrowing amendments. 

Let's take another example on a different issue of added matter, where the issue also came up only after a communication of intention to grant issued. Again, this relates to a divisional application, which had a fairly easy ride through examination and a communication of intention to grant issued with claim 1 in the following form:

In this case it is more difficult to see immediately what the issue is but, on analysing the claim features in relation to what was filed as the parent application (see below for a link to the register), it is clear that there were significant added matter issues. Third party observations were filed that pointed out this to the examiner, stating the following:

The observations were properly and clearly argued, and raised substantive issues that were not raised during examination. They were therefore clearly relevant. One would expect that this would result in examination being reopened and the A76 issues addressed in full. This is, however, not what happened. Instead, the examiner provided the following explanation of why the observations were not relevant and did not justify examination being reopened: 

I would hazard a guess that any vaguely competent European patent attorney with even a limited experience of dealing with added matter objections at the EPO would be surprised at these comments. This is not something we would expect to see coming from an examiner at all, but sound more like something coming from a representative trying to justify some tenuous intermediate generalisations for a divisional application that did not have direct and unambiguous basis for the combination of feature as claimed. The result was that examination was not reopened and the application could then proceed to be granted with claim 1 in the form as allowed. 

The lesson from these examples (if they can be considered representative at all) seems to be that third party observations have to be pretty damning for an examiner to be willing to reopen examination. I wonder how clear an added matter objection would need to be for an examiner to consider observations to be sufficiently relevant to reopen examination? Is there a hidden threshold of relevance that a third party has to meet? The Guidelines do not explain. Am I perhaps unfairly making a general rule out of an unusual example, or is this a more common occurrence? As always, comments are welcome.

For those wishing to investigate further, the applications in question are published as EP4092953 (double patenting) and EP4117228 (added matter).