by Dennis Crouch
The chart below provides one view of prior art citations in utility patents over the past decade or so (Jan 2005 – May 2018). I divided up cited references into three broad categories: Top) Non-Patent Literature; Mid) US Patent Application Publications; and Low) All other US & Foreign Patent Documents. The chart shows per-patent averages for each of these categories grouped by patent issue year.
There are two easy descriptive conclusions to draw from the cart: (1) Overall, the average number of cited references has continued to rise during this time period; and (2) while each category has grown, most of the growth is in the citation of US patent applications.
The USPTO only began publishing applications in 2001 and so growth was expected. Published patent applications are especially useful because their timing of publication (18 months after filing) corresponds well with the prosecution timeline (First action at approximately 15 months). For patent applications that never issue as patents, the publication does more than simply shift-back timing — those patent documents would have never entered the public domain. I have not yet gone back to research what percentage of cited published applications fall into this second category of published-but-never-patented.
For 2018 (through May 31), the prize-winner is APPLE’s U.S. Patent No. 9,986,419 with 6,701 citations (six thousand seven hundred and one)! The patent claims “social reminders:”
Claim 31. A method comprising:
at an electronic device:
receiving input representing user instruction to provide a reminder in the future, the user instruction identifying an entity; and
after receiving the input:
detecting, by a microphone of the electronic device, an audio input;
identifying, in the detected audio input, a voice corresponding to the entity; and
in response to identifying the voice, providing the reminder.
The image above shows the way it would work. User asks iPhone to “remind me to ask Joe when I see him;” When iPhone hears Joe’s voice nearby, it provides the reminder to User. I’ll note here that the patent was issued without rejection.
Coming in a close 2nd and 3rd in terms of most-prior-art-citations is Apple’s U.S. Patent No. 9,966060 (User-Adapted Speech Synthesis and Recognition); and Apple’s U.S. Patent No. 9,865,248 (Text to Speech) — both of which have 6,000+ cited references. (In fact of the patents issued thus far in 2018, the 8 with the most cited references are all Apple patents.
I called Apple the “prize winner” since they submitted more prior art references than any other entity or inventor. The problem – of course – is that it is impossible for the examiner to consider all of these references, And I am willing to bet that neither the patent attorney nor the inventors read all of these. In an article I highlighted last year, Prof. Jeffrey Kuhn termed any patent with 250+ citations as “impossible” — because the examiner simply would not have time to review those citations before making a decision on patentability. Kuhn wrote that a “small proportion of patents [are] flooding the patent office with an overwhelming number of references.”
“A method comprising: at an electronic device: receiving input representing user instruction to provide a reminder in the future, the user instruction identifying an entity; and after receiving the input: detecting, by a microphone of the electronic device, an audio input; identifying, in the detected audio input, a voice corresponding to the entity; and in response to identifying the voice, providing the reminder.”
“Hey Mabel, if you hear Rich out there at some point today, remind me to step out of my office and tell him about…”
At the risk of sounding Moonbeamish, this seems like an archetype for “do it [it=something we’ve done forever] on a computer” type claiming.
What are the scope of claims?
Any possible manner of setting-and-getting such reminders on any electronic device?
Or just the manner delineated in the patent (assuming any manner of doing so was set out with particularity)?
Wow…. no “human mind” rejection? No abstract idea rejection? Who was the examiner, Michael Cohen?
My hats off to the prosecutor.
In 1972 I authored an article published in JPOS defining a new form of patent fraud- citing a voluminous list of prior art so extensive that the most relevant piece of prior art is buried in the cited references. I know it was cited in at least 1 Court decision where such an issue arose.
Michael, I think you may be referring to the old Penn Yan Boat case? But as far as I am aware the Fed. Cir. itself has not yet specifically held that burying a material reference in the middle of a large pile of other citations is IC?
“burying” appears to be a pejorative term.
As long as there is a consistent (and internally consistent) manner of reporting references (not limited – and not preclusive either – of large sets of references), any such “accusation” of burying should be an invitation into the accuser’s motives.
Um, examiners and prior art searchers preferentially cite to PG-Pubs because they have easy to identify numbered Paragraphs. Much easier to say “See Paragraph [0065] noting…” than to manually try and figure out what Column / Line you want to cite.
Yes, that sounds trivial, but it’s a few minutes per citation and increasing efficiency equates to dollar signs.
I can think of three reasons to cite to the USPTO, namely i) duty of candor defensive (CFR Rule 56, is it?) ii) to bolster sufficiency of disclosure iii) to set the claimed subject matter in the context of the relevant state of the art. Only the last of these three bites, at the EPO. I assume it is overwhelmingly the first of these three which applies, at the USPTO.
Whether to cite to the USPTO the early-published A publication or the much later published B might depend upon the reason for its citing to the USPTO. For i) who cares. For ii) and iii) one ought to cite the A publication, surely, because it is the thing whose purpose is to put the skilled person, as early as possible, in a position to enjoy (and benefit from) its teachings.
“it is the thing whose purpose is to put the skilled person, as early as possible, in a position to enjoy (and benefit from) its teachings.”
That’s strictly a view of the government’s action to publish, and has no bearing on the choice of which to use – which bespeaks of having a choice.
Your innate bias shows by you jumping to the wrong topic…
input representing user instruction to provide a reminder in the future
Totally different from “input” representing user instruction to ignore something in the future! The latter is very volatile and has a completely different fake structure.
Just ask a Judge Lourie or Judge Newman! They are very serious people.
Your feelings, and rant, are noted.
Would be interesting to see breakdowns in Pre-Grant publications: those that eventually did and eventually did not result in a grant.
If there is a patent and a pre-grant publication, I will cite the later, simply because citing to paragraph number is easier than column/line numbers.
And I realized you mentioned that you haven’t done that work yet, I just wanted to point out one reason to cite a pre-grant publication over the corresponding patent.
Thanks PC.
Curious if the data include all utility patents or just those with citations?
I was reading some time ago that some larger filers were considering to forego IDS filing in divs/cons and instead rely on the IDS filings in the parent (except of course if new art should come to the attention of the applicant – not sure how often that happens so far downstream).
All utility patents. However more than 99% include some cites. 10-20% have no “applicant cites.”
Thanks Dennis.
Easy enough to check, the patent at number one is indeed a continuation from (also granted) patent US 9,668,121.
Terminal disclaimer in child.
Perhaps more interesting is the fact that there are only nine pages of Figures and 13 pages of specification.
That being said, the parent DID “survive” one Non-Final rejection, and made two sets of amendments (the second set by way of Examiner amendment). ALL submitted items were signed off as being “considered.”
Will a patent or published application like “the prize-winner .. APPLE’s U.S. Patent No. 9,986,419 with 6,701 citations” be of defensive protection value against later-filed patents of others on the subject? That is, with all of that in one document, will it be easier to make a 103 defense argument against the unobviousness of someone else later claiming combinations of features from those commonly cited references?
Does an inventor or attorney need to read an entire reference to believe that it might be material? Isn’t the duty of the inventor/attorney one of assessing materiality — not whether the examiner has sufficient time to review? Does the CFR say anything about the examiner having sufficient time to review as a constraint on submission?
This. As an attorney, if the Applicant or assignee tells me that a reference is material, then I will cite it. I am under no duty that I am aware of to confirm their understanding, since both they and I are under the duty to disclose.
There IS no such duty (even as examiners and others may attempt to view the situation otherwise).
Heck, there is not even a duty to perform ANY search…(albeit I will quickly add that such is most definitely not a part of Best Practices).
Note that “social reminders”, “user-adapted speech synthesis and recognition”, and “text to speech” are all related, and are likely internally designated in the same “family”. Did the inventors and/or patent attorneys read all 6000+ references in connection with this application? Almost certainly not… But did they come across those references in drafting or prosecution of some of the related family members? Probably yes, hence how they got into the attorneys’ IDS preparation tool.
Given that 37 CFR 1.56(a) requires disclosure of “all information known to that individual to be material to patentability,” arguably, those attorneys have to cite every reference that came up during prosecution of those family members… and certainly better safe than sorry – not disclosing anything would inevitably result in some future litigator arguing to a jury that “they intentionally failed to disclose some reference, cited in a related application, because they knew that it would render these patents invalid.”
1) “hence how they got into the attorneys’ IDS preparation tool.”
I was thinking about this and wondered if there is any substantive difference between citing an application and citing a patent (that the application may have matured into).
Given the fact that there is a “no new matter” rule, one might conclude that the citation of an application suffices – for all time – for the “making the Office aware” aspect of why a citation in an IDS is being made. As the citation alerts the Office not only to the published item, but also to the entire image file wrapper, one might conclude that once something makes it into the “the attorneys’ IDS preparation tool” as an application, that there is no need for any later “clean-up” and changing application numbers in that tool to (matured into) patent numbers.
Does anyone see anything wrong with this line of thinking?
This would then tend to make the trend even more prominent for the growth of the Professor’s “middle” section, as the only additional growth to the bottom section would be foreign and those US applications that had exercised a non-publication request.
.
2) Separately, one has to wonder at all those items that did not employ a non-publication request and (for a variety of reasons**) did not mature into a granted patent: a fundamental imbalance may then exist in so far as the two sides to the Quid Pro Quo deal are out of alignment in such circumstances.
**the variety of reasons most definitely includes some cases where the sharing does not deserve patent protection because of the 102 and (limited) 103 reasons that the items provided by applicant’s are already known – as practiced; but the converse also holds true (and reflects the recent attempt at saying “these kinds of innovation and advances do not belong in the patent system”) that these other reasons mean that the government has obtained its end of the Quid Pro Quo bargain, published that gained aspect of the bargain, and the other side got nothing.
For this (and other) reason, non-publication should be the default, and one should need to affirmatively select publication in order to more truly balance (or perhaps more accurately, return to balance) the Quid Pro Quo.
One arguable reason to cite the issued patent and not just its published application is if they are by the same applicant as the subject application and the claims have been so amended as to present a double patenting issue?
As to selecting U.S. non-publication of applications as the default, that of course means giving up any patent protection for that invention in the far larger total market that is now outside of the U.S. Also, it precludes the earlier product-defensive statutory bar of the published application. [If the published application is applied by examiners, that can save what would otherwise require an IPR, interference, or other litigation against later-filed applications of others that issued as patents before yours.]
Paul I will (attempt) to forego the usual snark, as at least one of your replies has a glimmer of meaningfulness.
As to your first offering, I think that the designation has zero effect on any actual eventual double patenting issue, and the double patenting – whether it exists or not – is quite apparently going to exist if one goes down either path of reporting on an IDS applications or granted patents.
As to your second offering (the one with a glimmer), it goes without saying that in those case wherein non-publication is not available, the default would be abrogated. Further, since you apparently have not been involved in prosecution for quite some time, non -publication requests are NOT fixed in stone, and IF any subsequent action does arise (for example, later deciding to apply in a foreign sovereign), the non-publication request can be rescinded without any dire effect. Yes, this is something that would have to be evaluated, but would be no different than – and thus may be concomitant with – any such foreign filing decision.
As to your third point, well, this returns us (yet again) to the point that I have YET to see a cogent reply concerning – the “decision” of the Office to entirely forego a meaningful and not insubstantial source of KNOWN (albeit secret) prior art. Let’s just move along here, unless someone can actually come up with a cogent reply.
But I do thank you for attempting to provide an answer, even if you come up short.
Yes, I was aware of being able to withdraw a non-publication request, but the deadline to avoid losing Paris Convention priority for foreign filing as well as the U.S. application is too short to be useful. Note the last sentence here: “35 U.S.C. 122 *****(b) PUBLICATION.—*****(2) EXCEPTIONS.—*****(B)
(i) If an applicant makes a request upon filing, certifying that the invention disclosed in the application has not and will not be the subject of an application filed in another country, or under a multilateral international agreement, that requires publication of applications 18 months after filing, the application shall not be published as provided in paragraph (1).
(ii) An applicant may rescind a request made under clause (i) at any time.
(iii) An applicant who has made a request under clause (i) but who subsequently files, in a foreign country or under a multilateral international agreement specified in clause (i), an application directed to the invention disclosed in the application filed in the Patent and Trademark Office, shall notify the Director of such filing not later than 45 days after the date of the filing of such foreign or international application. A failure of the applicant to provide such notice within the prescribed period shall result in the application being regarded as abandoned.”
? Your purported “too short” is not – especially as I indicated that any and all such decisions to file foreign are made at the same time of the rescinding of the non-publication request (as should happen anyway). One just shouldn’t do the one without – at the same time – doing the other, so any time buffer is total extra.
Comments are closed.