The Mythology of Prior Art
The mythology of prior art is that prior art is something that prevents a patent being issued.
There is a simple moral logic behind this:
- The prior art existed before the patent applicant "invented" the invention.
- Therefore the patent applicant cannot claim to have invented the prior art.
- Therefore the applicant cannot be granted a patent on the "invention", unless the patent is re-written to both acknowledge the prior art and to delete any claims that are no longer sufficiently novel given the existence of the prior art.
For a typical repetition of this myth, here is an extract from an article on IPWatchdog by Gene Quinn, patent attorney:
In each of these three cases we would say that the earlier reference of knowledge is prior art that prevents a patent from now issuing.
- If the invention in question was described in a patent issued anywhere in the world prior to the patent applicant inventing it, then no patent can be obtained.
- If the invention in question was described in a printed publication published anywhere in the world prior to the patent applicant inventing it, then no patent can be obtained.
- If the invention were publicly known in the US, but not necessarily patented or published, prior to the patent applicant inventing it, then no patent can be obtained.
Gene even uses the word "prevent" (emphasis is mine).
What's wrong with this myth is that prior art doesn't "prevent" anything. Prior art is not an active agent. Prior art is documents, and patents, and sometimes web pages. All of which are passive things. Nothing gets prevented until someone finds the prior art. And not just anyone – the patent examiner examining the patent in question has to be made aware of the prior art, and has to be made aware of it before the patent is granted. Then, and only then, can prior art "prevent" anything from happening.
The problem with the myth is that it makes the patent system seem fairer than it really is, because it tells us to believe that patents are never granted on prior art. But actually patents can be granted on prior art, and they are granted on prior art. And more importantly, even after this fact is discovered, a patent granted on prior art is not automatically invalidated. Nor is there any straightforward affordable procedure for initiating such a invalidation. It's like if you stole something, and no one noticed until after it was declared to be yours, and then you get to keep it.
How To Get A Patent on Your Patent's Prior Art
In the following scenario, prior art exists, but the prior art does not "prevent" a patent being granted:
- A non-patenting inventor invents an invention.
- The non-patenting inventor publishes the invention in some manner that counts as "public disclosure".
- The patenting inventor re-invents the invention. (He or she may or may not have read the original public disclosure. We will probably never know.)
- The patenting inventor applies for a patent on the invention.
- The patent examiner searches for prior art.
- The patent examiner does not find any prior art.
- The patent is granted.
- The non-patenting original inventor becomes aware that the invention has been patented by someone else.
After all these events have occurred, anyone attempting to implement the invention as described in the publicly disclosed prior art will be infringing on the patent granted to the patenting "inventor".
But Can't Prior Art Be Used To Invalidate a Patent After It's Granted?
Yes, there are means by which prior art can be used to invalidate a granted patent.
But, none of these means are cheap, or easy. For example:
- A "third party" can request a patent re-examination. In the United States, the minimum fee for this appears to be US$6000.
- An implementor can implement the invention, and then wait to be sued by the patenting "inventor" for infringement. And then, in court, the prior art can be raised as a defense and as a reason for invalidating the patent.
Note: Invalidation and partial invalidation
In the example above I assumed that the prior art actually is the patented "invention", such that, in theory, the prior art would completely prevent the patent being granted, if it became known to the patent examiner, or, if the patent is challenged after granting, the prior art would result in the patent being completely invalidated.
In practice such exact equality between prior art and a patented invention rarely occurs. More likely is that the prior art and the patent partly overlap, which results in the patent being partly prevented (pre-grant), or partly invalidated (post-grant).
But this does not alter the substance of the arguments I give here – whether or not prior art completely or only partially invalidates a patent, the existing system fails to sufficiently "protect" existing prior art from future patents.
Second-Class Citizenry for Unpatented Prior Art
The patent system supposedly reflects a "bargain", where "disclosure" is traded for a limited period, i.e. 20 years, of exclusive rights.
In theory these "disclosures" are of significant value. (Although some would argue, that in some fields, for example software development, "disclosure" is virtually worthless.)
Yet the patent system does not provide for any independent formal recognition of disclosures which are not part of a patent application. If "disclosure" was really considered to be a major public benefit, then every disclosure should be considered worthy of official public recognition, whether or not the disclosed invention interferes in some way with the validity of a patent, or a patent application, and whether or not such a possible patent application has yet occurred.
There is no public mechanism by which an inventor can submit, to a patent office, or any other public institution, prior art, as prior art.
Prior art can only be submitted as prior art for something, i.e. for a patent.
In other words, the patents are the first-class citizens, and any unpatented prior art inventions are the second-class citizens that only exist in relation to the patents for which they are prior art.
And if the submitter doesn't have deep pockets, prior art can only be submitted as prior art for a patent which has been applied for but which has not yet been granted.
It follows from this, that if you are aware of a publicly disclosed invention (be it your own or someone else's), and you want to make sure that it does not get patented, you have to constantly do patent searches, in all patent jurisdictions (or at least in those that matter), looking for all patents for which the publicly disclosed invention might be prior art. Just to make sure that you can submit the invention as prior art within the correct time window to the relevant patent examiner of the relevant patent application.
Which is a lot of work, just to make sure that something which should be free actually stays free.
Fame and Obscurity: Why Tim Berners-Lee and the World-Wide-Web is a Misleading Example
You've probably heard the story about how Tim Berners-Lee invented the World-Wide-Web and decided not to patent it, so now we all have the web for free without any restrictions imposed by Sir Tim.
Certainly no-one has attempted to patent the World-Wide Web after Tim Berners-Lee invented it, although there have been a few claims to have invented it before he invented it. (And those claims have been robustly challenged with the assistance of multiple interested parties.)
Knowing this story about the invention of the web might inspire you to publicly disclose your invention, setting it free for future generations.
Yet it is quite possible that publicly disclosing your invention won't guarantee the freedom of everyone else to use your invention.
In the first instance, it is very unlikely that your invention will get to be as well known as Tim Berner-Lee's invention. Especially if you lack the resources to start implementing it straight away. (Tim Berners-Lee was given the go-ahead by his boss to start creating an implementation working in his existing job – he didn't have to quit his job or raise funds or remortgage his house or anything like that just to get started. And he was working for an organisation that apparently did not require its employees to patent everything that they invented.)
But when it comes time for the patent examiner to examine a patent application, if your prior art invention has not been successfuly implemented and come into widespread use, then very likely your supposedly publicly disclosed invention will be completely unknown to that examiner. It will be as if it didn't exist at all.
And you probably won't get to hear about the patent for which you have prior art until the patent is granted. At which time it will be too late to easily and cheaply challenge the patent.
Moving Slightly in the Direction of Giving More Priority to Prior Art ...
USPTO Encourages Third Parties to Participate in
Review of Pending Patent Applications
This press release tells us that the
Ask Patents website was created
"utilizing input from the USPTO".
This is a slight improvement over what has been previously available, in that it provides a well-known location to which anyone can post prior art for successfully granted patents, without any requirement to pay for it.
I have taken advantage of this new opportunity myself by asking a question CALL FOR PRIOR ART: (Twitter) "Device independent message distribution platform" (8401009) and then answering my own question.
This is a step in the right direction, but it is still very patent-oriented. The patent or patent application has to be known to exist before one can submit prior art. One is not encouraged to ask questions such as: "CALL FOR PATENT FOR WHICH THIS IS PRIOR ART: 'Miski: A White Paper'".
And "Ask Patents" is not a legally defined institution. It is a private website, and there is no legal requirement for any data submitted to it to be retained indefinitely or to be included in any official public record.
A Proposal: Formal Registration of Publicly Disclosed Inventions
What we need is a prior art system that disconnects the submission of prior art from any hypothetical future patent or patent application that the "art" in question is prior to.
Such a system would best be described not in terms of "prior art", but rather in terms of "public disclosure".
It would work very simply as follows:
- The government, of whatever country, would create a formal registration system.
- Any person or organisation could submit a publicly disclosed invention.
- Each disclosure would contain the following elements:
- The identify of the registrant
- An indication of who "invented" the invention (if known)
- A title
- Content of the disclosure, formatted in some well-known standard format (i.e. plain text, maybe HTML)
- A pointer to any existing evidence that the disclosure was made prior to registration (for example a Wayback entry)
- The content would necessarily be licensed by the submitter for free re-distribution (ideally, as is the case for patents, content should be put into the public domain).
- Following registration, the registration system would issue a unique ID, and then make the registered disclosure available via a standard URL which would include the ID.
- For the purpose of aggregating and verifying all disclosures made over a particular period of time, it would be desirable to calculate and retain a cryptographic hash of the disclosure.
A system in which the government undertakes to publish indefinitely any content submitted by any person is open to various types of abuse, so some basic restrictions would be applied, including:
- Submitters should be formally identified by some means.
- There should be a maximum length to any individual disclosure.
- Submitted disclosures should be reviewed to check for content that may be offensive, libelous or unsuitable for any other reason to be part of a publicly available government publication.
- A moderate fee should be payable, to reflect the costs of accepting, reviewing and publishing each disclosure on the web. (An optional extra fee might pay for inclusion in a regular printed publication of submitted public disclosures.)
Direct benefits
By itself, such a registration system does not guarantee that patents will not be granted on prior art.
What it does do is provide strong and affordable protection to publicly disclosed inventions. Once an invention is publicly disclosed and registered, it will not "disappear off the map". It will not be subject to the vagaries of public web archives. And for any such invention, it will be slightly harder for patent applicants to claim "I didn't know about it".
International agreements
As far as I know, nothing in existing patent law prevents the establishment of a registration system for publicly disclosed inventions, and presumably nothing in any international intellectual property treaty prevents it either.
So any government that wants to set up such a registration system can set it up.
And if one government sets up a registration system for public disclosed inventions, and it is seen to be a good thing, other governments will set up their own registration systems.
Possible long term effect on the patent system
Anything that "protects" prior art will help to mitigate some of the badness that comes out of the patent system. But it will not completely eliminate the threat that patents pose to human freedom in general. Almost by definition, every patent granted is a corresponding reduction in human freedom, until such time as the patent expires.
But a registration system for publicly disclosed inventions would be an institution whose raison d'être is to protect public freedom to use those ideas which deserve to be free.
The existence of such a system can only help to highlight the value of freedom in general. As such, it may prompt more people to starting asking themselves questions like: "Are we all better off with a system that rewards inventors by giving them the right to reduce the freedoms of other people?"
Which I think would be a good thing.
Appendix: Some Prior Art on Systems for Public Disclosure
Statutory Invention Registrations
Until very recently the United States has had a system of Statutory Invention Registrations.
The option of registering these "statutory" inventions has now been abolished. It would appear that the requirements for such a registration were:
- Write out the description of the invention as if it was a patent
- Pay at least US$920
It is perhaps not surprising, given these requirements, that only 2264 were ever issued in 26 years.
(Reference: Brown & Michaels, also thanks to Maynard Chang who mentioned this type of disclosure in his comment.)
ip.com
ip.com is a commercial website that will publish your public disclosure for only US$225, and publish it in a database likely to be searched by patent examiners.
"Antipatents"
In Transparent Patents, Carl Malamud of media.org talks about, among other things, antipatents, where an "antipatent" is a public disclosure of a non-patented invention (although he seems to be describing an organised non-governmental disclosure system, and not one that is formally and legally defined).