Are Applicants ALWAYS Able To Survive Art Unit 2138?

Art Unit 2138 has issued patent applications for "Computer Architecture, Software, and Information Security: Memory Access and Control" at a clip of 87.5% (1626 out of 1859 cases) post-Alice.

That said, are applicants ALWAYS able to survive AU 2138? Not quite. 4.4% (82 out of 1859) of AU 2138 cases finally disposed post-Alice had at least one Request for Continued Examination (RCE) but were still abandoned. Yes, that means 95.6% of cases in AU 2138 were either issued, or the applicant gave up in the first round of prosecution.

Taking a slightly different approach in comparison to my previous post on how to successfully navigate AU 3689, I hope to explain or at least hypothesize in this post how the remaining 4.4% of applications became abandoned in AU 2138 post-Alice. Where did it go wrong? Here are some data-driven tips to avoid the identified pitfalls.

1. Be Willing To Continue Examination In AU 2138

As suggested, 95.6% of cases in AU 2138 were either issued, or the applicant gave up in the first round of prosecution. 64.8% (151) of the abandoned applications (233) in AU 2138 did not have any RCE filings. In other words, the applicant was not willing to continue prosecution, perhaps based on financial considerations or perhaps because the applicant was unaware of the statistics for AU 2138.

To be fair, some of the abandoned cases without RCEs may still have been abandoned after RCEs were filed. Only 25.2% (410) of the grants required even one (or more) RCE, and the average number of RCEs filed per granted case was 0.34. If you consider only those cases where at least one RCE was filed, your overall allowance rate in AU 2138 drops from 87.5% overall to 83.3% (410 out of 492) if you reach the first RCE. I think most financially able applicants would take an 83.3% allowance rate, considering it still exceeds the average at the USPTO, even for newly filed cases.

2. Avoid Written Description, Enablement, and Indefiniteness Issues Under 35 U.S.C. 112

93.7% of cases in AU 2138 were either issued or had at least one rejection under 35 U.S.C. 112 ("112 rejection"). 49.4% (115) of the abandoned applications in AU 2138 had at least one 112 rejection, with 18.5% (43) of the abandoned applications having at least one 112(a) rejection (lack of written description/enablement), 39.5% (92) having at least one rejection 112(b) rejection (indefinite claim(s)), and 8.6% (20) having both. In comparison, only 26.8% (435) of the issued applications had at least one 112 rejection (8.9% (144) with 112(a), 21.0% (342) with 112(b), and 3.1% with both).

In other words, applicants who abandoned cases in AU 2138 frequently did not adequately describe, enable, and clearly claim their inventions. This issue is usually avoidable by careful patent drafting, and USPTO applicants generally do avoid this issue. As explained by James Cosgrove in 2017, 112(a) rejections are avoided by applicants 88.3% of the time across the USPTO, and 112(b) rejections are avoided 78.8% of the time.

It should be particularly easy to avoid 112 rejections in the first action since the application of 35 U.S.C. 112 is not dependent on any prior art findings. Of the 115 abandoned applications with 112 rejections in AU 2138, a staggering 62.6% (72) of these (30.9% of the total abandonments in AU 2138) had 112 rejections in the first action. The second tip is for the applicant to take advantage of the ability to freely and carefully craft the specification and claims to steer clear of the predictable application of 35 U.S.C. 112 pre-filing.

3. Do Both Tip #1 And Tip #2

98.8% (1837) of cases in AU 2138 were issued, did not have an RCE, or had a 112 rejection, leaving only 1.2% of cases without any of these characteristics. In other words, only 9.4% (22) of the abandoned applications included at least one RCE without any 112 rejections. Does that mean applicants are always able to wear down examiners in AU 2138 for well-written specifications and carefully crafted claims? Still not quite, but what happened to these 22 remaining cases?

4. Be Willing To Pay The Issue Fee

Of the 22 (1.2% of the total) remaining cases with at least one RCE and no 112 rejections, two were actually allowed by AU 2138. The applicant simply did not pay the issue fee after the examiner sent the Notice of Allowance.

5. Don't Try To Patent The Same Thing More Than Once

Of the 20 (1.1% of the total) remaining non-allowed cases with at least one RCE and no 112 rejections, 6 received an obviousness-type double patenting rejection on the first office action. This means that the claims in the abandoned case were likely not patentably distinct from claims that had already issued in another patent owned by the applicant. In other words, the applicant already got the relevant patent rights.

6. Be Careful With Foreign Priority Documents And Provisional Applications

Of the 14 (0.8% of the total) remaining never-allowed cases with at least one RCE and no 112 rejections, 6 depended from an application filed outside of the U.S. (5) or a provisional application filed within the U.S. (1). As explained in my post about provisional application quality and the corresponding Patently-O guest post, 26.7% of U.S. provisional applications are missing 25% or more of the unique word stems in the broadest independent claim of the first-filed child non-provisional application. The translation issues stemming from priority cases filed in a different language pose a more difficult hurdle for applicants to establish a good priority claim.

If you do need to use a foreign priority document or a provisional application, my suggestion to further improve your odds is to be careful when preparing the priority document and in the translation process. All that said, English is my only fluent language, and I did not look into the quality of the priority documents or the translations.

7. Do Not Pursue Unreasonably Short Claims

Of the 8 (0.4% of the total) remaining never-allowed cases with at least one RCE, no 112 rejections, and no potential provisional or foreign priority issues, 4 were filed with claims of less than 75 words in length. As explained in my post about extremely short patent claims and the corresponding IPWatchdog post, there is roughly a 0% chance of allowance for a software-related patent application claim of less than 75 words. Not only that, the strategy of filing extremely short patent claims in the software space results in longer prosecution and narrower claims.

8. Avoid 112 Issues That Steer Your Claims

Three out of the four remaining cases had 112 issues that did not show up as 112 rejections. These issues likely steered the claims and arguments into uncomfortable territory for the applicant.

Case A had an interview summary immediately prior to abandonment where the examiner objected to a proposed amendment as not being supported by the specification (i.e., on 112 grounds even though it was not a 112 rejection). See Tip #2.

Case B had an interview summary immediately prior to abandonment where the examiner objected to an argument that hinged on a limitation that was not stated in the specification (i.e., on 112 grounds even though it was not a 112 rejection). See Tip #2.

Case C had an initial objection to the specification for not describing the claimed subject matter (i.e., on 112 grounds even though it was not a 112 rejection). See Tip #2. Also, the examiner (same as for Case D) had a 61% allowance rate, which was the third lowest allowance rate for the art unit.

9. For The 1 Out Of 1859 Cases, Potentially Reconsider Your Strategy For Examiners Who Have A Significantly Lower Allowance Rate

Case D had a corresponding PCT application that was abandoned before US prosecution began, potentially signaling a change in the importance of Case D to the applicant. Case D was abandoned after the fourth office action with only one RCE in prosecution, where a reference initially used as 102(b) prior art became part of a 103 rejection. (Notice this is the first time I mentioned any issues relating to the prior art.) The examiner (same as for Case C) had a 61% allowance rate, which was the third lowest allowance rate for the art unit.

I look forward to your comments below! Thanks again to LexisNexis PatentAdvisor® and Innography® for providing the relevant data through their powerful tools.


  1. Eric,

    Your approach of a sort of counter-factual to your 3689 analysis is an interesting one. After reading it several times, I had the following thoughts in no particular order (my apologies if these stray from the intent of the post):

    To me, your post really reinforces the need to be careful with 112 in complex areas of technology more than anything else . An art unit with a high allowance rate would indicate that overcoming the prior art is relatively easy (duh) so these problems become magnified. I'm curious as to how this analysis might play out in a strictly mechanical art unit with a similar allowance rate.

    Along the same lines, it'd be interesting to see in how many of the abandoned applications (especially those without RCEs) have a 112 in the last OA before abandonment. I think that this is important because 1st action 112(b) rejections are often times simple to overcome whereas in my opinion a 112(a) rejection is more likely to be fatal to an application, and the 112(a) doesn't typically surface until an amendment is made that adds new matter. It would also be neat to know how many have ONLY 112s, and how many had 112s and also prior art rejections in the last OA.

    In line with your comments regarding claim length, it'd be interesting to see the length of the applications themselves (allowed vs abandoned). With 112, a relative lack of length suggests a lack of detail and/or fall-back embodiments. This makes it more likely to run into fatal 112(a) and/or (b) issues due to lack of disclosure or insufficient detail, and also more difficult to overcome prior art.

    In addition to the reasons you suggest that might explain reasons why an RCE wasn't filed in this AU, the low number of RCEs in this art unit would seem to indicate that this is an AU with examiners and applicants that would tend to know the art pretty well and already have an idea of what exists in the art - they cut to the chase right away. I suspect that the number of first-action allowances is probably pretty high. Related to this, it also would explain the drop in allowance rate post-RCE - in such an art unit where the issues tend to get settled early, it would seem that the longer you go into prosecution the more likely it is that the Examiner's position regarding denying patentability will become hardened.

    I'm not sure what to make of the apps that you pointed out re: priority, especially since those are of applications that did NOT get 112 rejections. Did those have prior art rejections where the priority document might have made a difference? It'd be interesting to see if the abandonments with 112 with foreign priority are in line with the domestic priority/no priority 112 abandonments.
    Translation issues can be dicey in computer/software because there's much less in terms of figures to fall back on as disclosure.

    Finally, given the high post-Alice allowance rate and the fact you didn't mention it, it would seem that 101 wasn't really an issue. It'd be interesting to see what the pre-Alice allowance rate was - my guess is not that different.

    Hopefully I didn't stray too far from the point of your post, but going through your analysis got my mind going in several different directions, and I enjoyed the process!

    Keep the posts coming!

    -Tomas Prieto

    1. Thanks for the thoughtful comment, Tomas! I hope to keep adding fresh data-driven and tool-driven perspectives to the patent practice.

      Alice was and still is almost a non-issue in AU 2138. Without date filtering, out of 1534 final office actions analyzed by PatentAdvisor for AU 2138, 83 (5%) had a 101 rejection. In contrast, this is almost the only issue for AU 3689.

      It's hard to know/say whether "examiners ... know the art pretty well" if that art never determined the outcome of prosecution. I think in any group of people, some are more diligent than others, and attentive supervision tends to help with diligence. I know the supervisors are active in AU 3689 because they're often pulled into the interviews. I'm not as familiar with AU 2138.


Post a Comment