There is no copyright in a file format ... unless there is
There's a lot of litigation around Bitcoin at the moment, on a number of different fronts. The litigation around the purported fiduciary duty of developers is interesting, but for another day.
Today, we're looking at file formats and copyrightability, following last week's decision in Wright & Ors v BTC Core & Ors [2023] EWHC 222 (Ch).
The interesting bit of the case is whether or not copyright arises in the so-called "Bitcoin File Format".
The issue of copyrightability of file formats is far from new, and the decision of Mellor J is worth reading if only for a recap of the case law so far on this point.
The "Bitcoin File Format"
Mellor J summarised the Bitcoin File Format in the following (lengthy) terms:
There is no doubt as to what the Bitcoin File Format is. It is described in Schedule 2 to the Particulars of Claim. The overall structure of a block comprises three parts: 1) a block header of 80 bytes; (2) the vtx number, of 1-9 bytes, which records the number of transactions in a variable VarInt; and (3) the transactions recorded in the block, of variable size.
The block header is always 80 bytes but Schedule 2 explains that it comprises various data fields as follows:
i) nVersion, 4 bytes in length comprising a 32 bit unsigned integer stored in little endian and representing the version of the block format;
ii) hashPrevBlock, 32 bytes in length, being a double hash of the previous block header stored in little endian – essentially the link to the previous block in the chain;
iii) hashMerkleRoot, again 32 bytes in length, comprising a double hash of the root of the Merkle tree of transactions stored in little endian;
iv) nTime, a 4 byte field, representing the time in seconds since 1 Jan 1970, in a 32 but unsigned integer in little endian;
v) nBits, a 4 byte field recording a measure of the difficulty target in compact format;
vi) nNonce, a 4 byte field. Nonce is a portmanteau of 'number used only once'. The Nonce is used in the proof of work algorithm and is effectively the number that blockchain miners are solving for.
No copyright in the Bitcoin File Format...
Based on what the Bitcoin File Format is, the claimants always had an uphill struggle in proving that copyright existed in it, in my opinion.
Mellor J seemingly felt the same way.
He found - or, at least, did not dispute - that the developer "expended substantial skill and judgement in creating the Bitcoin File Format, such that the originality/intellectual creativity requirement is met".
But he - rightly, in my view, following SAS No. 3 - held that there is a distinction between an intellectual creation and a copyright work. There are more requirements to establishing copyright than something merely being the product of intellectual creation.
The bar for copyright is already very low, and the court was not included to further lower it.
The sticking point for Mellor J was that the Bitcoin File Format was neither expressed nor "fixed" anywhere.
In simple terms, the requirement of "fixation" is that the work must be fixed in a tangible medium of expression. In other words, documented somewhere, in a permanent form.
He said:
It remains the case that no relevant 'work' has been identified containing content which defines the structure of the Bitcoin File Format.
and
[T]here is no evidence that the Bitcoin File Format is set out in any part of the software or early blocks written to the Bitcoin Blockchain, as opposed to the Bitcoin Software simply reading and writing files in that format.
and
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.
In the absence of fixation, there was no copyright work, he ruled.
The claimants asked for permission to appeal, which the High Court turned down. The Court of Appeal may feel differently.
... but perhaps in other file formats?
In his detailed considerations of previous cases, Mellor J opined - obiter, since XML formats were not at issue in this case - that some file formats, particular XML formats, might be copyright works, protected as literary works (not computer programs).
He said:
Some (particularly XML file formats) contain sufficient content (and not just structure) to sustain a claim to literary copyright. Others (perhaps the majority) may not.
Citing Technomed Ltd v Bluecrest Health Screening Ltd [2017] EWHC 2142, Mellor J said:
It is therefore clear that copyright subsisted in the 'XML Format' because it contained 'content – not just structure'. Note that the content is not the 'text from the Classifications, the Options and the Patient Definitions' ... which is included in a particular XML file. Instead, it is the terminology used in definitions (self-describing using natural language terminology) of the data which are to be included.
I must admit that I struggle, conceptually, with the notion that copyright should restrict an XML format, even one with cleverly-constructed definitions, but the law is not currently with me on that.
Points to take away
If you are looking to reuse someone else's file format, you'll need to consider whether, in doing so, you are infringing copyright. The borderlines as to whether something is copyrighted, or copyrightable, are blurred rather than sharp, so this could be a challenging exercise. One might focus on content, rather than structure, and then for more technical legal matters such as fixation.
And if you are looking to protect your file format, the emphasis - at least, unless this decision is varied on appeal - would seem to be the basics of copyright: being able to demonstrate that the work is the product is original / a product of your own intellectual creation, that you can demonstrate fixation, and that the work is the kind of work in which copyright can arise.