10 Secrets to Locating Non-Patent Prior Art

Guest Post by Stuart Soffer, IPriori, Inc.

1. Finding prior art is a 'degrees of separation problem': you are separated from your art by some number of people and connections.

2. Build a timeline

Timelines aid visualizing the evolution of technology and locating the sweet spot of your prior art. Build multiple timelines in parallel (looking like a sheet of music) each line tracking a separate aspect. Possible lines to include are: the prosecution history of each patent; an industry timeline, i.e., the sequence of releases of Microsoft Windows operating systems; corporate history and accused product development; and prior art as you assemble it.

3. Get on the phone

Don't fear cold-calling. If the patent identifies people, or papers cited look promising, locate and contact the authors to see what materials they maintain, or if they can refer you to others. One search led me to the cell phone number of a CEO, was on a golf course when I called – he was very helpful pointing me to the right person.

4. Find the packrats and hoarders – those that build their own collections, either from their career, or interest. These folks exist, but they don't publicize their collections, it isn't indexed, and they may not appreciate visibility. Other sources are from estate sales and antique stores, but this is better for proactively building a collection for future use.

5. A by-product of the non-patent prior art search is identification of potential testifying experts. Sign 'em up.

6. Accused 'infringers' could have their own prior art

Long established companies, especially those with formal research groups and product evolution, will possibly have their own prior art to current products.

7. Days and weeks matter

Relevant prior art dated mere days or weeks after the effective filing date for a patent is frustrating. This is instructive for parties contemplating filing patent applications that the time you delay filing can come back to haunt later on with prior art dated a few days earlier. There are instances where the prior art is just a short time after the effective filing date. In one instance the relevant non-patent art was an article in a conference proceeding. Presumably publication date is the conference date. That date didn't predate the priority date, however I researched the date submittals were due for peer review. That date was before the priority date and, under the right circumstances, that can count as prior art. 

8. Establish a company historian

One search some years ago led me to the AT&T Corporate Historian. This was an actual position.

9. Multiple path searchers and searchers; don't rely on only one modality. Different searchers don't get identical results. They come to the problem with different personality, background, preferences in search sources and search techniques. An important search will use more than one searcher or company to get better coverage.

10.Always be on the lookout for prior art

Some searchers continue a search under the understanding the likelihood of similar requests in the future. Be opportunistic: visit museums or antique stores as you travel. Have a camera to take pictures of artifacts you encounter to aid in recall (easy to do with smartphones.)

11.(Bonus) Be aware domain taxonomies, such as Library of Congress Subject Headings, Medical Subject Headings (MeSH), ACM Computing Reviews Categories. Build a vocabulary of synonyms with which to vary search queries. Here are some samples:

  • Memory, storage, disk, array, RAM, DRAM, flash
  • Signal, indicator, message, bit, semaphore, indicia, flag
  • Display, window, CRT, VT100, terminal emulator
  • Connected, attached, communicates with    
  • Module, program, server, layer, client, abstraction, applet, application

53 thoughts on “10 Secrets to Locating Non-Patent Prior Art

  1. Your dripping sarcasm is nothing new – it’s part of you being an @_$$.

    Now, if you actually had anything smart to add, then there might be some nominal humor in your posts.

    As it is, there is not.

  2. that Congress needs to amend Title 35 yet again?

    You are STILL in way too much of a hurry to tr011.

    Try again, and this time read what I actually posted.

  3. After reviewing the narrative and the Prior Art, the Applicant will write the original claims.

    Interesting proposal, but it is the claims that should provide focus to the examiner. Otherwise, the Examiner may go on a wild goose hunt on a technology that was discussed in the specification for enablement purposes only.

    The current system works best when (i) Applicant presents claims of varying scope (from broad to narrow) and (ii) the Examiner fully searches those claims. However, if the examiner finds art that could be easily amended around (e.g., it is from another field and a simple limitation can get around it), then the Examiner needs to find better art.

  4. That’s because you have sux searchers apparently.

    Tell me 6, how many searchers are there in the private world that specialize in your art unit and your art unit alone? The best searchers for any art unit should be the examiners within the art unit itself — that is what they do every day. Nobody else can say that.

    What did I write earlier:
    This is the Examiner’s greatest contribution to the entire process — knowing the art

  5. Unless you care what the resulting patent is worth.

    What is that supposed to mean?

    Most career inventors are reasonably well acquainted with the state of their art.

    No … in most cases, too much art is out there. Plus, they have no clue about art in different fields that could be brought in with a 103 rejection.

    The problem is generally that the person drafting the claims is less familiar with the art, and tries to draft nice, broad claims that the applicant can look at and say “yes, we’d love a patent on that!”

    What’s the problem with that? If the claim is too broad, then the prior art will require an amendment. If the attorney did his/her job well, they’ll be dependent claims of varying scope that will be on the correct side of the line that divides patentable from unpatentable.

    And then you wonder why you get rejections that aren’t based on the “best” art. Why should the examiner spend his time searching for the best, if your claims are broad enough that any reference from the past year (or ten) will do?

    Ah … because that NEVER happens (at least to me). I take that back, out of the thousand plus applications I have prosecuted, I have seen maybe 2 or 3 applications with claims so broadly written (not by me) that it was exceedingly easy to find prior art.

    I’m talking about the citations of art that require heavily-strained reading to apply to the claim language (i.e., the classic “well, the language is somewhat close, so I’ll cite it under 102 and call it a day).

  6. You can’t have the world’s experts employed at the office, because then they can’t get patents.

    There is a difference between being an expert in developing new technology and being an expert on the state of the pre-existing technology.

    You become an expert on the state of the pre-existing technology by continually researching it — i.e., what an examiner should be doing every day.

  7. What the suggestion here does is offer a path of obtaining prior art – from those supposedly best at doing so, that is, examiners – with the added benefit of how the Office is likely to look at the application (BRI-style), WITHOUT the current adversarial and actual patent-impacting events.

    I see… so my suggestion that any applicant can do exactly that right now if he wants to … and that doing it within the PTO after filing (as proposed) causes no end of drafting and administrative problems … quite misses the point. Please accept my most sarcastic apologies.

    Let’s focus on the point, then. What part of this zany scheme is so important, and so inaccessible to today’s applicant, that Congress needs to amend Title 35 yet again?

  8. You quite miss the point IANAE.

    What the suggestion here does is offer a path of obtaining prior art – from those supposedly best at doing so, that is, examiners – with the added benefit of how the Office is likely to look at the application (BRI-style), WITHOUT the current adversarial and actual patent-impacting events.

    You do know that applicants are NOT REQUIRED to do ANY prior art searching at all, right?

    Sure, an appliant can (and should) engage in prior art searching as a best-practice, but the idea advanced here goes beyond that.

    Your too-quick-to-tr011, smart-@_$$ without the smart ways get in the way (yet again). Thanks for proving my point (again) about where the real low-quality blogging comes from.

    I suggest you look for a Calvinball-Facespike-Addiction program.

  9. Proposal:

    The rules currently permit the applicant to do exactly what you propose. It’s called searching the prior art before you file. Still want an early filing date? Throw in a single claim you don’t expect to prosecute (e.g. Diehr’s claim), and replace it later. But why would you ever commit to a spec before having some idea of what you think the claims should look like?

    If you’re filing an application, and you can’t draft an independent claim covering an “inventive concept” that you reasonably think is patentable, you’re doing it wrong.

    As for the idea that you can tell the examiner of your filed patent application what to search, and that it won’t end up in the file wrapper… I can’t think of a better way to sum up how very wrong that is than this:

    anon: “I see value in your idea.”

  10. Interesting.

    How about break the service cleanly away from patent prosecution to eliminate any later judge made estoppel effects?

    Offer the service in a completely “safe” environment (no publication to tip off competitor).

    No extra strings to later actually filed applications – the existing duty of candor would suffice to maintain applicants’ integrity.

    Have a way of priority claiming and partial priority claiming (similar to CIP).

    These ideas would go a long way to restoring an atmosphere of professional cooperation.

    I see value in your idea.

  11. Proposal:

    Write new rules that permit the Applicant to file a non-provisional patent application with a Spec, Drawings and NO claims. Additionally, the Applicant will file a list of Inventive Concepts (based on the Spec and Drawings) for which the Patent Examiner will search. (The examiner can also search for any other concepts disclosed in the Spec and Drawings but not listed in the Inventive Concepts.)

    (The new rules will expressly state and require that the list of Inventive Concepts will not be retained as part of the file wrapper.)

    Then, the examiner will search those concepts and find the best Prior Art. After the search, the USPTO will send out a brief Office Action that includes a short narrative discussing the best Prior Art references with respect to the subject matter disclosed in the Spec and Drawings.

    After reviewing the narrative and the Prior Art, the Applicant will write the original claims.

    It is my hope that this would cause the original claims to be more narrowly tailored towards patentable subject matter.

    In my opinion, if this were done, then both the USPTO and the Applicant would be able to prosecute claims much more efficiently (fewer RCEs).

  12. “Put simply, the applicant is LOCKED IN at the time of filing – there is nowhere to go.”

    Please explain this in more detail. When read in light of the previous statement (i.e., the Applicant has received a “high quality” Office Action), I don’t understand what you are trying to say.

    Are you saying the following: the Applicant can amend the claims only within the confines of the original Spec, Drawings and Claims?

  13. “the search training is not mandatory and not well compensated”

    This is a shame. I deal with both the EPO and the USPTO. The search results I get for US applications are very very… very far behind the search results I get for EP applications in terms of relevance. And this helps considerably the EPO examiners doing a proper job.

  14. You make my point by missing my point (yet again).

    Thanks 6 – anyone still wondering why we have an Examination problem at the Office?

  15. That date didn’t predate the priority date, however I researched the date submittals were due for peer review. That date was before the priority date and, under the right circumstances, that can count as prior art.

    There is normally a presumption of confidentiality when a paper is submitted: if it is accepted, it will be published at some date, otherwise the world shouldn’t hear about it, and the reviewer isn’t supposed to disclose or use the information (at least in theory). In EP-land you won’t get that far with late documents. But do check that the priority claim is valid!

    Try to find out who the inventors are, where they worked or studied previously, and look up their names in databases. They will sometimes do you a favor by publishing before the filing date. Even if there is a grace period you could be pleasantly surprised.

  16. Making 112 1st paragraph does its job in all cases is my job? Um, no, I don’t work on all cases, and I certainly am not capable of making sure that the board (or every examiner) has its head removed from its arse every time it makes a 112 1st decision.

    Try again ta rd.

  17. which is never going to happen.

    The point of that 6, is THAT is your job that is “never going to happen.”

    Thanks for making my point.

  18. Google “Ricardo Oltra” and “World Patent Information”. If you’re at work you can download/read the article. He’s one of the many EPO examiners that regularly write high quality articles on the science behind searching. Another person to Google in that magazine is “Evert Nijhof” who writes more readable articles.

    Btw, EAST is a tool created in the 90′s which makes it super dated. Maybe we’ll get the PE2E system by 2016, if we’re lucky.

  19. Nah, not really. You’d be surprised at how much “implicit” nonsense applicants can dream up and say was disclosed, or even just plain ol’ made up nonsense that they put into their claims. Then you’re depending on 112 1st to always do its job, which is never going to happen.

  20. “which no other database on EAST does”

    Just because east also sux mah wa ng doesn’t mean that Derwent doesn’t do the same.

    And yes, I’ve heard all about how awesome Derwent is. How come practically every time one of our searchers uses it in their “expert” capacity they come up with trash art? Because the whole thing is gar bage. That’s why.

    Not worth your time unless you’re scortching the earth and desparate.

    “but how else do you find Korean or Chinese art of relevance?”

    You hire chinese and korean searchers that you trust. That’s how.

    Or, you learn chinese and korean. Which, btw, it would be easier to learn japanese, korean, chinese, german, french, and spanish (practically the only people that practice and disclose in my art) than to use Derwent over a lifetime.

  21. “•Module, program, server, layer, client, abstraction, applet, application”

    I almost didn’t catch the joke here. program, abstraction, synonyms.

  22. 6 – As a salaried employee, if you don’t get paid (other time) for job training then you’re getting the shaft. And from what I hear, the training you do get is pretty subpar. I review examiner search strategies in the file wrappers all day, every day. None seem to have received any training beyond “Do a full text key word search.”

    Re: Derwent – Yes, it sucks compared to a fully text searchable English language document. But this data encompasses 20,000,000 pieces of prior art that you cannot search in that manner. To say it sucks misses the whole point. It provides a human generated abstract of non-English language documents, uses a unique indexing system, a whole separate classification system (separate from the CPC/ECLA/IPC/USPC systems), uses a standardized thesaurus of terms, and gives you family data which no other database on EAST does.

    Long story short, it doesn’t suck. It definitely has major issues that suck (can I get a simple dedup or more than a single image please), but how else do you find Korean or Chinese art of relevance?

  23. You rather missed the point 6.

    No doubt a larger application is more formidable to examine. None-the-less, all applications – no matter the size – are indeed locked in at time of filing.

  24. “However, I don’t see that rate being acheived.”

    That’s because you have sux searchers apparently. At least in my art I can find a 102 at the bare minimum for the independent claims, for at least 1 out of 4. Probably 2 out of 4.

  25. “Finding prior art is the area in which examiners receive the least amount of training.”

    We have weekly trainings on how to search, the STIC’s will help you hands on.

    Idk where you got this dream about them getting little search training. It is just that the search training is not mandatory and not well compensated.

  26. Let us say for a moment that I agree with you NWPA, and that what you have said is “astute”.

    What then can possibly be done about the situation? You can’t have the world’s experts employed at the office, because then they can’t get patents.

  27. “Put simply, the applicant is LOCKED IN at the time of filing – there is nowhere to go.”

    Not if they bothered themselves to write a longer spec, say 200 pages or so.

  28. reply to EG,
    When I was an organic chem graduate student, I used Beilstein extensively. There is a real treasure trove of citations to old chemical journals in Beilstein. As a patent attorney, I have prosecuted many organic chemistry patent applications in last 25 years and have yet to see a reference from Beilstein cited by an examiner.

  29. Unless you care what the resulting patent is worth.

    Non-sequitur.

    Most career inventors…

    Maybe you want to read what was actually posted and try again?

    get rejections that aren’t based on the “best” art

    Another non-sequitur. You ASSume that the examiner’s rejections “will do,” – that is an assumption not evident – and certainly not evident in this conversation.

    Or are you (yet again) “spinning” what other people are saying?

  30. Why not examine ALL the claims, like you’re supposed to.

    They typically do, which is why you get the other common complaint on this board – they allowed a claim, but I don’t want it because it’s narrower than claim 1.

    Doesn’t much matter anyway, if claims 4 or 5 are “fairly narrow” but still not inventive over claim 1.

  31. Claim 1 may be broad. Chances are claim 4 or 5 is fairly narrow. Why not examine ALL the claims, like you’re supposed to.

  32. Definitely not arguing with your economic guesstimate… In an ideal world, a prior art search would not eliminate a filing but would instead eliminate the first three Mickey Mouse (TR) rejections. Unfortunately that is not reality.

  33. a search needs to knock out perhaps 1 out of every 3 or 4 disclosures to be cost-effective.

    Unless you care what the resulting patent is worth.

    Most career inventors are reasonably well acquainted with the state of their art. Whatever they don’t know wouldn’t be published yet for a searcher to find. The problem is generally that the person drafting the claims is less familiar with the art, and tries to draft nice, broad claims that the applicant can look at and say “yes, we’d love a patent on that!”

    And then you wonder why you get rejections that aren’t based on the “best” art. Why should the examiner spend his time searching for the best, if your claims are broad enough that any reference from the past year (or ten) will do?

  34. And better still, if a little more time was invested in the prior art search pre-filing, thousands more could be saved.

    The problem is that inventors and the attorneys who draft the applications are not necessarily experts in the art or searching the art. Some inventors know the art, but (most) many do not. An examiner in a particular art unit will examine hundreds and review thousands of pieces of art in that art unit. This is the Examiner’s greatest contribution to the entire process — knowing the art.

    With the economy, more clients see a prior art search as they would undercoating at a car dealership.

    The value added just isn’t there. Just a guesstimate, but a search needs to knock out perhaps 1 out of every 3 or 4 disclosures to be cost-effective. However, I don’t see that rate being acheived.

  35. Superb list Stuart, you and IPriori are the leaders. My Top Three of your Top Eleven…

    #3 People rely too much on email and forget the power of a personal call.

    #4 Despite their own shows on basic cable, the collectors/hoarders are a great group and are usually eager to help.

    #10 It helps to be a geek and have an eye for the obscure. I’m always in old bookstores, The Strand, Powell’s, and the owners themselves can be great resources.

  36. “I would rather have an examiner give me a killer 102 reference in the first office action than in the fourth office action 2 years later.”

    Great, well said. And better still, if a little more time was invested in the prior art search pre-filing, thousands more could be saved. With the economy, more clients see a prior art search as they would undercoating at a car dealership.

  37. Finding prior art is the area in which examiners receive the least amount of training. There are heaps of memos on 101, 102, 103, 112, CAFC rulings, and Supreme Court cases but very little assistance and/or teaching about searching.

    To give a concrete example: The USPTO pays a hefty amount to integrate the Derwent WPI data into their system. But this value-added data requires searching in a manner different than searching full text US documents. The USPTO has the Derwent experts from TR teach examiners during their second week of hiring only. This is before they have a clue as to why they need the information being given to them.

    Another concrete example: We’ve rolled out the CPC at the USPTO. The world experts on the CPC are the instructors and trainers at the EPO. Guess how many of them have given lectures to the US examining corps? You know the sad answer as well as I do.

    BTW, good guest post. Searching NPL can be a bottomless pit and exhaustive searching of NPL should definitely be saved for the big litigation budget cases. Even a pro can charge you 10k, 20k, or 100k and hand you a blank sheet of paper for your results.

  38. For organic chemistry, my late patent attorney father used to look in Beilstein (a compendium of organic chemical synthesis in Germany from the late 19th Century) which had many, many organic synthses which weren’t to be found anywhere else. Might not be of as much value now, but you never know.

  39. The problem is anon is that the examiners often don’t understand the technology. You can’t search when you don’t know what you are looking for.

  40. +10000000

    And that’s the real underlying problem with the Office.

    As I have stated, both myself and my clients know a c_rrp rejection when we see one and will continue to fight.

    A solid rejection simply removes the B$ from the table. Sure, you will still have the (very very few) outliers who will simply not give up, but those are a piece of cake is the first office action is a high quality office action. Put simply, the applicant is LOCKED IN at the time of filing – there is nowhere to go.

    As soon as the office apologists understand this, we can dispense with the other agendas hiding under those skirts – the “oooh, patents are bad” agenda, the “oooh, software patents are bad” agenda, and the “oooh, functional claims in patents are bad” agenda.

    Do the job right – the first time.

  41. What separates a quality examiner from the average examiner is his/her ability to find good prior art.

    I’ll be the first to admit that it can be very difficult. However, I would rather have an examiner give me a killer 102 reference in the first office action than in the fourth office action 2 years later.

  42. “But, that is the curious thing about litigation. I mean, what use would it be for me to search and find one-touch shopping, or swipe to unlock?”

    Try to sell them to the opposing counsel? Make 1M$ and call it a day?

  43. 6′s 6 secrits to locating NPL:

    1. Start on google very broadly, find and build the vocabulary (and possibly list of important people in that sub-field) you will want to be using for the rest of the search. Find a few references at least near the subject matter you’re looking for (not necessarily having the limitations you’re looking for).

    2. Read the semi-relevant references you found with a quick skim to see if they mention anything about the subject you are interested in for your search, if they do, look for the citation to previous works that they will usually make.

    3. Look up those previous works cited to.

    4. Repeat from step 2 until you run out of citations to check.

    5. Repeat from step 1, narrowing your search as you run out of citations from each search and better understanding how a person of ordinary skill would word what you’re looking for.

    6. Profit. Usually after 2 or 3 iterations at least. When you become zen like 6, you will have your results faster.

  44. Good advice. Thanks.

    Two sources that I have found useful: (1) The German diplom ingenieur theses are often lengthy histories of the technology and they are written in English and often available on the Internet. I have also faced killer rejections from examiners with these.

    (2) Small conferences that are very specialized. The papers for those conferences are often detailed descriptions of hardware and software. I am sure, by the way, that I could knock out Apple’s patents. But, that is the curious thing about litigation. I mean, what use would it be for me to search and find one-touch shopping, or swipe to unlock? Both of which, I’d bet I could find from the early 1990′s in specialized conference papers.

Comments are closed.