Citing Patent References

by Dennis Crouch

011013_1244_1

The chart above shows a histogram of the number of references cited on the face of utility patents issued in the calendar year 2012.  As is apparent, this is a highly skewed dataset with a long tail.  Over the past five years, the percentage of patents citing more than 100 references has risen from 3% in 2005 to more than 8% today.  Obviously, an examiner is unlikely to thoroughly review more than 100 references during the ordinary patent prosecution process. Once we all agree on that fact, the 100+ references do not create real problem so long as the USPTO examiner has data processing tools to help discover important elements within that focused set of references. 

28 thoughts on “Citing Patent References

  1. “The IDS had about 200-300 listed prior art references, one of which was a patent from the ’70s that was directed towards a mechanical clutch pedal.”

    It was probably a typo. You need to keep in mind that there are practioners who think that their paralegals/support staff never, ever make a mistake and just blindly sign whatever is put in front of them. And then bill their client for a half hour of their time.

    Patents and applications directed to an industry standard are tricky. I’ve only once been involved in that and all I can say is I didn’t approach it any differently than my other applications. I counseled the client, including the engineers, the in-house counsel, the foreign associates, etc. about the duty of disclosure, reminded them every time I reported an OA, an allowance, a filed response, etc., and hoped they were complying.

  2. AAA JJ,

    Great movie – my favorite line is when Annie says to Nuke “would you rather I was sleeping in his bed and yelling out your name?”

    Great answer – i understand completely about having inventors not coming forth with relevant art until the prosecution is well down the road. With regard to my use of the word “diligence” I really didn’t mean anything more than a sanity check. One of the first applications i examined was directed towards a computer related invention. The IDS had about 200-300 listed prior art references, one of which was a patent from the ’70s that was directed towards a mechanical clutch pedal. No electronics or anything that would have remotely connected the thing to the claimed invention. You just have to roll your eyes and scratch your head on some of these IDS supplied references sometimes.

    One thing that I have worked into my examination practice over the years is to always ask the question “is this an area where there is likely to be some form of interoperability standard?” For some reason that is the one area where I don’t see a lot of submissions. In fact what is typical is to have the various entities that work on the standards turn around once the standard is published and then file patent applications that claim the same thing. There was a particular standard where a term was coined that was not common in the art prior to the release of the standard. I did a search on the term in the patent database using a date filter that filtered out everything prior to the release of the standard and as expected got zero hits. Then I changed the date filter to a date ninety days subsequent to the release of the standard and got close to 900 hits, all of which were applications filed between the release date of the standard and the following ninety days. I looked over a few of the applications and noticed that a good chunk of them were just straight copy-paste out of the standard. And yet nobody (at least in the applications I examined) ever listed the standard in the IDS as NPL. As the standard in question took about 3-4 years to draft, redraft etc. I would have to assume that these applications were being prepared while the standard was being worked out and then filed once the final version was released. Of the applications I examined that were related to this standard the patent database only supplied art that would yield a 103 rejection; however the standard itself was a clear 102(a) rejection on all of the applications that came across my desk. And frankly I was a bit puzzled as to why all these entities would file applications that were clearly anticipated by an industry standard. But there it was, plain as day.

    As a general rule I would say that the best references I see in an IDS are those supplied in examination by foreign offices; however even in those I see some that are dead on while others are a bit off target.

    i guess this is all just a hazard of the trade.

  3. Leopold,

    No, I can’t say that any come to mind. So having said that there would always be a risk that one of the references submitted in a large IDS might not be viewed as relevant during a text search perhaps because the terminology might be somewhat different than what is typically employed in the art or for whatever reason and then later turn out to be a relevant reference. Unless you happen to read each reference in toto it would likely never be discovered except through serendipity.
    I should add that I did have something like this happen recently where I was examining a continuation application and it turned out that one of the references cited in a 100 reference IDS did read on both the continuation and the allowed parent. The terminology in the reference was sufficiently different that it would not have been obvious to search for that expression. But it was very clear that it was describing the same thing as the application that I was examining. Frankly I think there will always be some outliers in a handful of applications that may read on the application.

  4. I did have a Crash Davis moment with your question (“Come on meat…”).

    Just kidding.

    I’m not aware of a single case of inequitable conduct being found for the case of a “dump” of documents where one, or more, of the documents renders the claim invalid. If there is one I’m formally soliciting the help of all of the legal gurus who populate this site to provide it.

    Not sure what you mean by “diligence” (and I’m not being snarky, the concept can be somewhat nebulous). Mr. Moatz, the former director of OED, insisted that applicants/practitioners had to read every reference in its entirety before submitting it in an IDS. I, and most (99+%?) rational people, regarded that as a rather stretched/strained reading of 37 CFR 11.18.

    Believe it or not, satisfying the Rule 56 duty is not easy. One of the engineers I work with sent me a spreadsheet of references related to a technology that he’s been keeping for years. It’s not a small number of references. He’s been keeping the spreadsheet for a couple years, and he inherited it from an engineer who retired after a 30+ year career.

    Okay, so now every time I file an app for this engineer, do I have to cite every reference on the spreadsheet? Do I have to read, and re-read, every single one every single time I draft a claim for one of his applications (he’s an inventor or co-inventor of at least 8 cases on my docket ready for drafting)? And do it again every time there’s a revision to any one of the claims in the applications I’m drafting for him?

    I’m in-house, so the time spent doing that won’t get billed, but still the amount of time reading, and re-reading, would make drafting an application for him prohibitively time consuming, and expensive if I had to enlist outside counsel to draft it.

    So how much diligence is required of me now that the knowledge of all these references has been imputed to me by the simple act of receiving an e-mail with the spreadsheet attached? Honesly, I don’t know. I don’t think anybody knows. I can only tell you that I’m actually trying to draft claims that are as broad as we’re entitled to and that I’m not simply throwing stuff up against the wall to see what sticks. For the company I work for, there’s simply no benefit in that.

    Of course, I’m not saying that doesn’t occur. But I’ve worked at a quite a few places, and have worked with literally a couple hundred+ patent attorneys in my career and I can tell you that I can’t recall a single instance of a colleague asking me for advice on how to circumvent the rules, or even suspecting that one of my colleagues was circumventing the rules. Again, not saying it doesn’t happen, but I’m not of the opinion, like some of the charlatans parading around on this site, that I’m the only ethical practitioner and that every other practitioner is a sleaze ball who would do or say anything to obtain an allowance.

    As to why one would wait to submit 100+ references, it’s possible that they were brought to the applicant’s/agent’s attention late, either by the applicant/assignee/3rd party. What do I do when I’ve told an inventor, who has been counseled over and over about the duty of disclosure, that we’ve gotten a notice of allowance and the inventor says to me, “Oh, great, by the way, would this be a good time to mention the spreadsheet I’ve been keeping for the past 20 years?”

    It happens.

  5. Hey plurality of thresholding units, from the years of posts you’ve been around the PTO for a while now, you still don’t know about Reference Manager in the eDan menu?

  6. I was thinking of some of the articles I have seen with regard to submitting a large number of references and perhaps triggering inequitable conduct.

    Have you seen any articles that suggest this as a likely outcome under current law? Absent a smoking gun memo where the patentee says that he is intentionally burying a relevant reference, I don’t see it happening.

  7. AAA JJ,

    Sorry about that – that came across like a hanging curve ball. Of course it would be an invalid claim. I was thinking of some of the articles I have seen with regard to submitting a large number of references and perhaps triggering inequitable conduct. Is there any diligence required on the part of the applicant and their representative to screen the references and provide guidance as to which might be the most pertinent or is throwing a large number of references without any regard to pertinence an acceptable practice? I’m just asking here, not slinging arrows.
    And as an aside, why wait until allowance to submit 100+ references in an IDS? Again, just asking…

  8. “Someday the USPTO will perhaps implement some of these features on their internal system.”

    First though they have to waste a couple hundred million dollars.

  9. “what are the implications of for example having an application allowed, filing an IDS with 100 references via an RCE, getting a signoff on the IDS and later having it come to light that one of the citations actually read on the claim?”

    If the citation anticipates or renders obvious the claim, the claim is invalid.

  10. The PTO could not require that.

    They do not have the authority to make applicants perform a prior art search and insert such estoppel type information (note that I am not saying that this might be desirable – just that the PTO does not have that authority).

    They can (and have) created optional paths of examination that do have such or similar mechanisms, but those are merely optional paths – not required ones.

  11. Anon,

    One of the features in most high-end search engines (not the free stuff online or on a US examiner’s desk) is the ability to limit the F&B citation searching to only X references from the EPO or references cited by the US examiner rather than all references. This significantly reduces the noise. Someday the USPTO will perhaps implement some of these features on their internal system. The presentation a few weeks ago of the new search tool showed some promise, but it is of course a long way off from being rolled out…

  12. PTO could require a 112 phrase like

    There are two candidates for the closest prior art. Physically it is probably the …

    Functionally the closest prior art is …

  13. For examiners and patent searchers, the large stacks of references submitted via an IDS weakens one of the usual tools for locating prior art. Classification has already been weakened by the hiring binge of the last decade combined with inadequate training. Text or key word searching can be problematic in areas without a defined nomenclature (all ‘hi-tech’ areas) and only goes back to 1976.

    Citation analysis (forward and backwards on references of particular relevance) used to help patch the holes in the above two strategies, now it generates too high a volume of references and must be limited by text or classification. This causes it to lose the independence from those limiters that it once provided. What this means in the real world (as opposed to the legal world of this blog) is that searching performed by examiners or others become just a little bit worse across the board.

  14. I should add that the Reference Manager tool allows you to get search strings that can be copied into our search engine with all the cited publication numbers from 892 and 1449 submissions. So once you have the whole collection in a string you can then combine that with text expressions to filter the collection.
    Now the question I have is this – what are the implications of for example having an application allowed, filing an IDS with 100 references via an RCE, getting a signoff on the IDS and later having it come to light that one of the citations actually read on the claim?

  15. Actually it can be done with Patents and PG Pubs through the use of our Reference Manager tool that is part of our Docket Manager. As far as NPL I think you have to get past the point of having paper submissions – otherwise the only option is to print the document to a PDF, open the document and OCR it, then search. Now you could do that for a whole collection of documents – Acrobat allows you to search a collection of dcouments in a file directory for example and I have done that for a limited amount of submissions. Many times this would be overkill though – if you have 20 documents that are submitted as NPL 3-5 might be published papers of length 6-20 pages while the rest may simply be screenshots of a browser page.

  16. There is a much easier way to do that. Go into eDAN, and go to any item in the file wrapper (the claims for example). Then go to the top, under tools go to “Reference Manager’ and a browser window will open with all the 1449 and 892 references that have been filed for an application. That way you can get the Patents and PG Pubs that have been cited as a BRS string that can be copied into EAST or WEST.
    Sadly, no such tool for foreign references

  17. “Obviously, an examiner is unlikely to thoroughly review more than 100 references during the ordinary patent prosecution process.”

    There is no requirement that the examiner give applicant submitted prior art any more consideration than the other 100′s, if not 1,000′s, of prior references that the examiner would “consider” during their search. See MPEP whatever.

  18. “We currently don’t have any search tools for searching the collections of NPL or foreign-language patent refs submitted on IDSes.”

    Considering the amount of money the CIO’s office of the PTO has spent on “IT” in the past decade or so, this is simply disgusting.

    What’s more disgusting is that the PTO spend hundreds of millions, if not over a billion, dollars on the useless first generation EFS (that probably had less than 10,000 applications filed), and on PAIR and all of the other barely useful and functional systems, and the people who wasted that money are all still there wasting money. Nobody got sacked for doing a lousy job. In fact, I’d be willing to bet they all got reviews that concluded they were doing record breaking outstanding quality work. And bonuses.

    In terms of usability and functionality, the EP Register is an iPhone 5 or Samsung Galaxy S III, while PAIR is a Motorola StarTac from the late 90′s. Disgraceful.

  19. That's right, the vast majority of the variance is created by patent applicants.  Examiners usually cite a handful of references and rarely cite more than a dozen references. 

  20. It’s interesting how VERY skewed the distribution is, with the mode (most frequent number of references cited) at 10 being less than one-fourth the mean at 43 (total #refs cited/total #patents – I would not say that the “average patent cited 43 references”). Looks like an income distribution, with the dishwashers at one end and Bill Gates and the bankers at the other. Dennis, is there any way you can parse the citations by applicant-cited versus examiner-cited in whatever tool you are using? They are indicated separately on the face of the patent. I expect you may find that the number of examiner citations is much less variable than the number of applicant citations – it’s hard to believe any examiner would have the time to write an office action with 100 references.

  21. We currently don’t have any search tools for searching the collections of NPL or foreign-language patent refs submitted on IDSes.

    They’re working on something to help out with importing lists of US patent documents from electronically submitted IDSes into the search tool, but currently, we have to OCR the IDS form, copy and paste the results into the search tool, and hope it gets it right; or else type them all in manually.

  22. Paul – Surprisingly to me, the PTO does not have a good tool for limiting a search to only the set of submitted documents — especially when the submitted documents are non-patent literature. So, my conclusion of “no problem” is merely hypothetical.

  23. Paul,

    What is the basis for your asterisk?

    What is the actual number and is that actual number actually surprising?

  24. What is the basis for the assumption here that: “..the 100+ references do not create real problem so long as the USPTO examiner has data processing tools to help discover important elements within that focused set of references.” ?

    Many of these applications with big “reference dumps”* include large numbers of technical publications, not just patents. Does the PTO provide a simple “cut and paste” system for putting all those paper citations into a search engine? [Obviously examiners do not have time to type them all in.] Do the PTO search engine databases have access to all those technical publications? What about the many claimed features that are not aminable to simple word searchs or are called different things in different references or need picture searches?

    *a surprising number of which seem to get sued on.

Comments are closed.