(Update: after this video conference and after this draft letter got circulated, the USPTO published an April 28, 2022 Federal Register notice that forced a substantial revision of this letter. The letter has now been signed and has bee sent to the Acting Commissioner. You can read about it here.)
Yesterday I had a productive video conference with Acting Commissioner for Patents Andrew Faile. The conference was encouraging in the sense that he seems to understand that the USPTO has completely bollixed up its handling of its DOCX initiative thus far. It was even a little bit encouraging in that he was grudgingly prepared to allow filers to provide a protective document along with the DOCX document, this of course being just what WIPO had been doing for ten years now with its “pre-conversion” document.
I was disappointed, however, to get a “deer in the headlights” reaction when I asked if he was familiar with WIPO’s “pre-conversion” document procedure under PCT Administrative Instructions § 706. It is as if the notion of allowing the filer to include, say, a protective PDF file along with the DOCX file was some new idea that had only just popped up within recent weeks or months. I had been telling lots of USPTO people about this for years now. The Seventy Three Practitioners explicitly recommended this in their letter dated September 27, 2019. How could it be that neither he nor his two minions who were with us on the video call had any idea that for ten years now, a PCT filer who was filing a character-based PCT patent application was able to file a protective document?
The conference was profoundly discouraging for yet another reason. It was clear that Andy is in some alternate reality in which he thinks that things might go like this:
- Andy and some USPTO developers in charge of DOCX, and a bunch of patent practitioners, will get together and settle on a plan for filing a whole bunch of patent applications in the next few months.
- The idea is that for each of these patent applications, the filer will upload a DOCX file and a PDF file of their own making.
- For each of these patent applications, the USPTO will then use its (proprietary) DOCX rendering engine to generate its own PDF file from the DOCX file.
- Within a year, maybe sooner, it will be up to the filer to speak up if they manage to detect whatever errors the USPTO may have created in the USPTO’s PDF rendering.
After the expiration of some evaluation period (he is a bit vague about this, but I have a sense he is shooting for this to wrap up in time to permit him to keep the present January 1, 2023 start date in place for the $400 DOCX penalty) then I gather he has in mind the USPTO would announce that very few if any discrepancies had been found between the user-provided PDF file and the USPTO-generated PDF file. Linked with this announcement would be a related announcement that the opportunity to file a protective PDF along with the DOCX file would be ended, given the USPTO’s triumphal finding that there had been very few discrepancies found.
Note by the way that as a general matter nearly all of the subject patent applications would be secret during the entire time of the evaluation period. So it would be difficult or impossible for members of the public to independently evaluate any claims by the USPTO as to whether there really were very few discrepancies. It would likewise be difficult or impossible for members of the public to independently evaluate whether such discrepancies were big or small. Perhaps most disturbing is that it would be difficult to imagine that practitioners would be willing to take the personal professional liability risk of subjecting truly complicated patent applications (containing difficult math equations, chemical formulas, tables and the like) to such testing at all. Very likely the only patent applications that any practitioner would be willing to subject to these risks in such an evaluation program are the simplest of patent applications containing only the very simplest of text. This would mean that even if by the end of this evaluation period few discrepancies were found, it would likely mean little or nothing.
It is also unclear how confident one can be that the applicants and/or practitioners doing the filings in this evaluation effort will be able to devote the time required to carry out top-to-bottom character-by-character proofreading of each PDF file involved within the limited number of months available. I know from my own practice that I am sometimes nearly overwhelmed by existing client obligations and it is sometimes almost impossible for me to fit an optional or discretionary activity into my professional schedule. I have to assume that many other applicants and practitioners sometimes face similar challenges in their own lives.
Finally, it is foolish, to put it charitably, to think that any evaluation program, no matter how thorough, that ends on any particular date (say, January 1, 2023), even if it were to uncover a complete absence of discrepancies between user-provided PDFs and USPTO-rendered PDFs, would have any predictive value for events occurring thereafter. Microsoft is well known for adding proprietary and undocumented or poorly documented features to its software, including undocumented or poorly documented extensions to its version of DOCX formatted word processor files. This often leads to situations where, for example, a user of some other non-Microsoft word processor is simply unable to open a DOCX file created by a Microsoft word processor, or is able to open it but is unable to view it correctly or print it correctly. We are unaware of anything that would prevent the USPTO from making unannounced changes to its proprietary engine for rendering DOCX files into PDF files. Any such events, occurring after the end of such an evaluation program, would make meaningless any conclusions drawn from an absence of discrepancies found during that evaluation program.
As I say, it certainly seemed to me that Andy is in some alternative reality where he thinks that some limited-duration evaluation program could somehow permit later concluding that “everything is okay with DOCX”, and that the $400 penalty can be kept in place, and that once it was determined that “everything is okay” then people would not need to be able to file protective PDF files any more.
Somehow he needs to be dragged out of his alternative reality. He needs to understand that it will never be the case that “everything is okay with DOCX”. He needs to understand that it is flatly impossible for any limited-duration evaluation program to yield any conclusion about “everything being okay” with DOCX.
With this in mind, I have drafted up a proposed letter to him. Let’s allow about one week for comments and suggested changes (please post them to the EFS-Web listserv). And then once we have the letter finalized (as I say, in about a week) we can collect signatures and then send it to him.
Andrew I. Faile
Acting Commissioner for Patents
USPTO
Dear Acting Commissioner Faile:
This letter comes to you from NNN members of the EFS-Web listserv. The EFS-Web listserv is a group of users of the USPTO’s patent e-filing systems, and the group include over five hundred attorneys and agents registered to practice before the USPTO. The signers of this letter, directly or through their law firms or corporations, have collectively prosecuted more than MMM US patents to issuance in the past ten years and have collectively paid more than PPP in fees to the USPTO in the past ten years. Our chief communications goal in this letter is to help you and your colleagues at the USPTO correct recent missteps in USPTO’s DOCX initiative. We also hope to help the USPTO do better about receiving character-based information from its customers and making productive use of that character-based information.
To summarize the most urgent points very briefly:
- USPTO must scrap its present software that provides false information (false “message digests”) in patent e-filing acknowledgment receipts about documents supposedly uploaded by users, and must publicly confirm when it has done so.
- USPTO must commit to some date certain upon which, going forward, the message digests in USPTO’s patent e-filing acknowledgment receipts will truthfully reflect the actual documents uploaded by users.
- For any filing of a new US patent application where the application is being provided in character-base form (for example DOCX) at USPTO’s request or requirement, the USPTO must provide an opportunity for the applicant to provide contemporaneously a protective version of the application in a user-selected format (for example PDF). The acknowledgment receipt for this filing must provide truthful message digests for both the actual uploaded character-based file and the actual uploaded protective file.
- For any filing of a follow-on submission (subsequently filed document) where the document is being provided in character-base form (for example DOCX) at USPTO’s request or requirement, the USPTO must provide an opportunity for the applicant to provide contemporaneously a protective version of the document in a user-selected format (for example PDF). The acknowledgment receipt for this filing must provide truthful message digests for both the actual uploaded character-based file and the actual uploaded protective file.
- The USPTO must explicitly recognize that if the USPTO’s rendering or processing of a character-based file were to deviate from the corresponding protective file, such a deviation might go unnoticed by USPTO personnel and by the applicant until such time as a resulting patent were litigated. Thus, some USPTO procedure for handling corrections (for example by petition or certificate of correction) must be available to the applicant for at least a long enough time to cover the term of the resulting patent plus its statute-of-limitations period.
- Any USPTO procedure for handling corrections relating to USPTO’s rendering or processing of the character-based file as compared with the protective PDF (for example by petition or certificate of correction) must be free of any fee cost to the applicant or patent owner.
- The USPTO must scrap its attempts to force the filer to agree, during the course of an e-filing session, to the supposed authoritativeness or finality of any document developed by the USPTO during the e-filing session.
We now turn from the summary of the urgent points to a more detailed discussion.
We appreciate and share the USPTO’s goal of receiving communications from its customers in character-based format rather than image-based format, so that the character-based information provided by customers would auto-load into USPTO’s systems.
ePave. Many of us were among the beta testers of USPTO’s ePave system of two decades ago, which was USPTO’s initiative for e-filing of US patent applications in XML format. Some of us e-filed hundreds of XML-formatted US patent applications in USPTO’s ePave system as beta testers. We are aware that the USPTO eventually abandoned that initiative in favor of the EFS-Web system for e-filing of PDF-formatted patent applications.
ePCT. Many of us are users of WIPO’s ePCT system. The ePCT system is very well designed from both the Office point of view and from the user (applicant and practitioner) point of view. The ePCT system permits e-filing of PCT patent applications in PDF format and it permits e-filing of PCT patent applications in XML format. The ePCT system also permits the user to see the status and content of application files and permits the filing of subsequently filed documents. The ePCT system provides a well-designed system by which the user can update bibliographic data (“92bis changes”). The developers of the ePCT system were in close communication with users throughout the design and implementation phases of the system and the developers continue to be in close communication with users. The developers monitor and participate in the relevant online user communities including the PCT Listserv, and it is commonplace for the developers of the ePCT system to respond to bug reports and to fix bugs within mere days of the reports. The ePCT system does not merely carry out validations of user inputs to serve Office wishes but also carries out validations of user inputs to serve user needs and to protect users from undesired results. The patent filing user community has repeatedly urged the USPTO to take WIPO’s example in these areas and has continued to be disappointed at USPTO’s failure to do so.
The signers of this letter, directly or through their law firms or corporations, have collectively filed more than MMM PCT patent applications in the past ten years, including more than NNN PCT patent applications through the ePCT system. We have also filed PCT patent applications in XML format (character-based format) at the Receiving Office of the International Bureau and we are familiar with the PCT rules that allow a PCT applicant to take the protective measure of submitting a “pre-conversion” copy of the PCT application along with the character-based PCT patent application.
Many of the signers of this letter are not only the attorney or agent taking responsibility for a client’s patent application from a professional point of view but are also often the person who actually carries out the e-filing activity on a patent office’s e-filing system. As such, many of us personally make use of EFS-Web, PAIR, Patentcenter, and ePCT. Many of us see with our own eyes the results of bad design decisions by the USPTO e-filing system developers but also the bad policy decisions by the USPTO in its e-filing systems, and we see with our own eyes the stark contrast with good design decisions in ePCT and good policy decisions in the rules underlying ePCT.
USPTO fails to make use of characters provided by users now. It is very difficult for users to take seriously USPTO’s demands to receive patent applications in DOCX format given that the USPTO fails to make use of character-based information that users presently provide. A few examples will illustrate this.
Issue Fee transmittals. Applicants transmit Issue Fees to the USPTO using Form 85B. Among other things, this form communicates an Assignee Name and Assignee address, and an Attorney, Agent or Firm Name. The legacy approach for Form 85B was that it was completed by the filer on a typewriter and sent by postal mail to the USPTO, and Issue Branch personnel would hand-key the Assignee Name and Attorney, Agent or Firm Name into USPTO’s systems for printing on the front page of the patent.
About six years ago, the USPTO released its web-based Form 85B. In the web-based form, the filer is required to provide computer-readable characters in the form for the Assignee Name and the Attorney, Agent or Firm Name. But what actually happens within the USPTO is that the computer-readable characters are flattened into images, and the same Issue Branch personnel continue to hand-key the Assignee Name and Attorney, Agent or Firm Name from the web-based Forms 85B into USPTO’s systems for printing on the front page of the patent. The proof of this is that there are many issued US patents where the Attorney, agent or Firm Name provided by the filer in the web-based Form 85B is spelled correctly (for example “Oppedahl”) and yet the name on the front page of the issued US patent is spelled incorrectly (for example “Oppendahl”). As another example, in the web-based Form 85B an Assignee city was spelled correctly as “Radom” (a city in Poland) but on the front page of the issued US patent it was spelled incorrectly (“Random”). This shows that Issue Branch personnel continue to hand-key such information from the web-based Form 85B, failing to make character-based use of the actual computer-readable characters provided by the applicant.
Corrected ADSs. One of the ways that an applicant can provide updates to bibliographic data to the USPTO is by means of a web-based “Corrected ADS” form in EFS-Web. This web-based form draws upon the existing bibliographic data in the USPTO system and permits the applicant to indicate which fields are supposed to change and what the new data should be. The web-based form then generates a submission with suitable strikethroughs and underscores. All of the data entry by the applicant is character-based. None of the data entry by the applicant is image-based.
We are disappointed to learn that in USPTO’s internal processing, the web-based Corrected ADS is actually handled as an image. USPTO personnel actually carry out the updates by means of hand-keying the changes into USPTO’s systems. The USPTO fails to make any character-based use of the characters provided by the applicant.
The user community has repeatedly, and for many years, urged the USPTO to take WIPO’s example and to handle updates of bibliographic data the way that ePCT handles it, with “92bis actions”. The user community has repeatedly urged the USPTO to take note that with the ePCT workflow, the user’s request goes to a processing team at the International Bureau of WIPO, and then the user’s character-based inputs are able to auto-load into the WIPO systems. There is no hand-keying of the bibliographic data by WIPO personnel.
For character-based patent e-filing, WIPO does it better. When WIPO first introduced XML filing for PCT applications, some of the signers of this letter were among the very first to carry out such XML filings of PCT applications. Some of us were at the same time carrying out XML filings of US patent applications in USPTO’s ePave system. From the very beginning of WIPO’s initiative for XML filing of PCT applications, WIPO was sensitive to the concerns of applicants as to whether mistakes might occur in the rendering or processing of the XML version of a PCT patent application. From the very beginning of the initiative, WIPO provided rules so that an applicant could file a protective version of the patent application. In the PCT rules, this protective version is termed a “pre-conversion format” document. The applicant is not forced by the patent office to adopt any particular fonts or word processor formats, and is not, for example, required to use Microsoft Word or any particular variant of a DOCX format. The applicant is able to use whatever word processor the applicant prefers and trusts. The protective version filed by the applicant (the “pre-conversion format” document) is permitted to be in whatever document format the applicant trusts to be authoritative.
The PCT rules permit the applicant to make use of the pre-conversion format document to correct any rendering or processing problems with the XML version by following certain procedures at the International Bureau. The natural question is how much time is available to the applicant to detect such problems and to carry out such corrections? The Treaty and rules of the PCT give the International Bureau jurisdiction over the procedural aspects of a PCT application for only a limited time, namely until thirty months after the priority date. As a consequence, this is the amount of time available to the applicant for such corrections according to the relevant PCT administrative instructions. We wish the time available could be longer, but we recognize that the IB has only this limited thirty-month period within which to grant such corrective relief.
It will be noted that WIPO’s way of incentivizing the applicant to provide a character-based patent application (here, XML) is nice to applicants: WIPO offers a filing fee reduction (here, 100 Swiss Francs, or about $107 at current exchange rates). This contrasts with the approach of the USPTO, which is instead to try to bring about the desired applicant behavior by means of a penalty (here, $400). It will also be noted that the actual avoided cost to the USPTO when it does not have to carry out OCR (optical character recognition) on a patent application is, by USPTO’s own admission, only $3.15 per document.
In the over ten years during which such XML filings of PCT applications have been taking place, there has never been even a single report known to the signers of this letter of any applicant anywhere in the world having found the pre-conversion format protective filing to have fallen short of any need of the applicant to correct any problem with Office rendering or processing of the XML-formatted document.
From the very first days of USPTO’s DOCX initiative, the practitioner community repeatedly urged the USPTO to take WIPO’s example and to permit the applicant to file a protective version of the patent application along with the DOCX-formatted version of the patent application. It is disappointing that until now the USPTO has refused to do so.
USPTO’s first DOCX procedure. The first DOCX procedure started in 2016, in USPTO’s “eMod Text Pilot Program”. Some of the signers of this letter were participants in this program. The way that this program worked is that the filer would upload a DOCX file to USPTO’s e-filing system (at that time, EFS-Web). The USPTO system would then render the DOCX file into a PDF file using USPTO’s proprietary rendering engine, and would present the newly created PDF file to the filer. It was then up to the filer to scrutinize the PDF file from the top to the bottom to attempt to find whatever mistakes the USPTO’s rendering engine had made. This scrutiny had to be carried out under great time pressure. On the screen of the e-filing system was an adhesion contract which the filer was invited to enter into by clicking “submit” to e-file the patent application. The adhesion contract, if agreed to, adopted the USPTO’s newly created PDF file as the authoritative document for all purposes in subsequent processing through any issuance of a US patent.
One problem with this procedure is the unreasonableness of expecting the filer to successfully detect any and all rendering mistakes during such a high-pressure time, often in the limited number of hours or minutes remaining before midnight so as to obtain a same-day filing date. If the rendering engine used by the USPTO were open-source, then filers could know ahead of time what might go wrong in such rendering and indeed could test the rendering ahead of time. But the rendering engine was (and is, to this day) a proprietary “black box” and, most disturbingly, the USPTO could change the behavior of its black box at any time and without warning.
A related problem with this procedure is the problem of ethical and professional responsibility for the practitioner. Under USPTO’s rules the patent practitioner is required to obtain a signature of the inventor on a Declaration of Inventorship. The Declaration calls for the inventor to have reviewed the subject patent application as part of the signing process. This review will necessarily have been linked to the DOCX file that the practitioner later uploads to the USPTO e-filing system (at some time after the inventor has reviewed the patent application and has signed the Declaration). But then, according to USPTO’s procedure, the authoritative document ceases to be that DOCX file and suddenly the authoritative document is a new PDF file that the inventor has never seen. It might be suggested by the USPTO that the newly authoritative PDF file is supposedly substantively identical to the previously authoritative DOCX file, but USPTO’s own actions (in devising its adhesion contract language) belie any such suggestion by the USPTO. There would be no need to force the filer to review the PDF file and to force the filer to assent to an adhesion contract unless there is actually some risk that the PDF file will fail to be substantively identical to the DOCX file. From this it follows that the USPTO at least tacitly acknowledges that it is unable to offer any assurances that its proprietary rendering engine will avoid introducing substantive changes into the PDF file. From this it follows that the USPTO puts the practitioner into what may be charitably described as a difficult position when suggesting that the previously signed Declaration of Inventorship may be filed in the application.
This first DOCX procedure was unacceptable to the filing community. Many of the signers of this letter were also signers of the letter from Seventy-Three Patent Practitioners to Brendan Hourigan of the USPTO, dated September 17, 2019. In that letter we expressed many concerns about the manner in which the USPTO was handling its DOCX initiative and, in particular, the propriety of the USPTO’s proposed $400 penalty for failing to comply with USPTO’s proposed DOCX procedure. We also reported many issues resulting from the filing of applications in DOCX format.
USPTO lied on August 3, 2020. On August 3, 2020, the USPTO published a Federal Register notice entitled Setting and Adjusting Patent Fees During Fiscal Year 2020 (85 FR 46932). Long before that date, many of the undersigned had notified the USPTO of issues resulting from the filing of applications in DOCX format. In particular, the letter from the letter from Seventy-Three Patent Practitioners to Brendan Hourigan of the USPTO, dated September 17, 2019 notified the USPTO of such issues.
It was thus astonishing to see the following sentence at page 46958:
To date, the USPTO has not received notifications of any issues resulting from the filing of applications in DOCX format.
This statement by the USPTO in the August 3, 2020 Federal Register notice was false.
USPTO’s second DOCX procedure. On June 2, 2021 the USPTO announced a second DOCX procedure. This announcement, in a Federal Register notice, proposed that “the DOCX documents filed by applicants” would “be the authoritative document”. This procedure apparently provided at period of at least one year during which the USPTO would preserve the DOCX document filed by the applicant, and it was apparently contemplated that during that one year the applicant could make use of the DOCX document as a protective document in the event that USPTO’s rendering of the DOCX file were to introduce an error.
A first problem with this procedure is that one year is not long enough. USPTO needs to acknowledge that mistakes made by the USPTO in rendering a DOCX file might not be detected by Office personnel or by the applicant during that one year or even prior to issuance of a patent from the application. There needs to be a procedure for correcting Office mistakes at any time during the term of an issued patent or during the statute of limitations running thereafter. Such a procedure necessarily calls for preservation of the protective document (or, at the very least, its message digest) for this entire time, not merely for one year.
A second problem with this procedure is that it is not reasonable to force upon every applicant the use of a USPTO-defined DOCX file as the protective document. WIPO recognizes this by permitting the applicant to select whatever document format the applicant trusts for use as the “pre-conversion format” protective document. The filing community has repeatedly pointed out to the USPTO that there is no single “DOCX standard” and that you can take any particular DOCX file and open it with any of half a dozen word processors, and you will see any of half a dozen non-identical renderings on the screen or on a printout. Math formulas and chemical structures render non-identically, page breaks happen in different places, and one word processor might fail to even know how to render a particular font that some other word processor knows how to render. In particular, Microsoft is very well known for adding features to its implementation of DOCX that it does not document publicly and that necessarily lead to unpredictable results when such a Microsoft DOCX file is opened by a non-Microsoft word processor. The fact that USPTO refuses to disclose the code for its DOCX rendering engine, despite years of being asked to do so, only heightens the unreasonableness of trying to force filers to adopt a DOCX format as the format to be employed for the protective document in this procedure.
As it turns out, apparently the USPTO never actually implemented this second DOCX procedure as set forth in its June 2, 2021 Federal Register notice.
USPTO’s third DOCX procedure. It appears that the procedure that USPTO actually adopted on June 2, 2021 is not the procedure set forth in the Federal Register notice. The procedure actually adopted on June 2, 2021 appears to be approximately the following:
- The filer uploads a DOCX file (call this “DOCX1”).
- During the e-filing session, the USPTO carries out an alteration of the DOCX file, resulting in a USPTO-generated DOCX file (call this “DOCX2”). The DOCX2 file is generally non-identical to the DOCX1 file, and in particular it nearly always has a non-identical message digest (the SHA-512 hash). The DOCX2 file is often a file which, in its visual rendering on the computer screen, looks at least on a quick review very much like the visual rendering of the DOCX1 file. Many filers might not even notice that the DOCX2 file had been created by the USPTO and slipped into the process flow in the e-filing process in place of the DOCX1 file.
- When the filer clicks “submit”, the acknowledgment receipt lists the message digest of the DOCX2 file. It does not list the message digest of the DOCX1 file.
- Once the e-filing process is complete, although we suspect the DOCX1 file is preserved somewhere in USPTO’s systems in a place that is not accessible to the applicant, the only file that is visible in IFW is the DOCX2 file.
- At any later time, if the applicant were to try to use the DOCX file that the applicant actually uploaded as a protective document, the applicant would be unable to do so, for two reasons. First, the message digest for the DOCX file that the applicant actually uploaded is not present in the acknowledgment receipt. Second, the DOCX file that the applicant actually uploaded is not visible anywhere in IFW.
A first disappointing thing about this procedure is that the acknowledgment receipt states a falsehood. The acknowledgment receipt states that the message digest appearing in the acknowledgment receipt is the message digest of the DOCX file that the filer uploaded. In fact this is false. The message digest appearing in the acknowledgment receipt is the message digest of the USPTO-created DOCX2 file, resulting from whatever alterations the USPTO made during the e-filing session.
A second disappointing thing about this procedure is that it fails to match what the USPTO promised in the June 2, 2021 Federal Register notice. The FR notice promised that the filer would be able to upload a DOCX file and that the DOCX file would be the authoritative document. Instead, the actual behavior of the USPTO is to discard that DOCX file and to slip into the official file a different DOCX file that is not the DOCX file that the filer actually uploaded.
USPTO’s proposed fourth DOCX procedure. We thank Acting Commissioner Faile for the courtesy of a video meeting on March 23, 2022 with Carl Oppedahl, one of the members of the EFS-Web listserv and one of the signers of this letter. During the meeting, as we understand it, the USPTO proposed a fourth DOCX procedure. In this fourth procedure, a filer who submits a DOCX file at USPTO’s request or requirement would be given an opportunity to provide contemporaneously a protective version of the application in a PDF format. If the filer were to discover that the USPTO had introduced some error in its rendering of the DOCX file, the filer could point to the protective version of the application to get the USPTO’s error corrected.
We note that this is precisely the approach that WIPO has offered for more than a decade under its PCT Administrative Instructions § 706. We further note that the filing community has been urging the USPTO to adopt this “protective version” approach as used with the WIPO PCT procedure, for many years now. The letter of the Seventy-Three Patent Practitioners to Brendan Hourigan of the USPTO, dated September 17, 2019 specifically recommended this “protective version” approach.
Notwithstanding this, the USPTO has been hostile to the “protective version” approach throughout its history of its DOCX initiative, most recently in its June 2, 2021 Federal Register notice:
Applicants should not submit PDF versions they created when filing an application in DOCX, as they are unnecessary. If the applicant submits documents in DOCX along with PDF versions they created (not the auto-generated PDFs created by the USPTO), then the DOCX version will still be considered the source or evidentiary copy, and the applicant will be required to pay the non-DOCX surcharge fee.
In other words, as of June 2, 2021, the applicant that attempts to provide a protective PDF version of the patent application would be punished by having to pay the $400 penalty.
Nonetheless we are encouraged to hear that the USPTO is now apparently willing to relent and to follow WIPO’s example from a decade ago, providing for the filing a contemporaneous protective document.
We were, however, discouraged to hear that apparently what USPTO contemplates is that this procedure might only be offered for some very limited period of time. The impression given to Mr. Oppedahl is that USPTO people have in mind that this would only be needed for some months, or maybe for a year, and that if some time were to pass during which very few USPTO rendering errors were detected, and if eventually USPTO were to feel that it had made enough repairs to its rendering engine, then USPTO could drop the “protective filing” procedure. It seems the USPTO people have in mind that eventually the “protective filing” procedure would be cut off and that filers would once again be left at the mercy of whatever proprietary DOCX rendering engine USPTO happened to be using at that future time.
It is very important that the USPTO understand that the USPTO has no choice but to pay attention to the actual needs of applicants and practitioners on these issues. USPTO’s proposed approach, as we understand it to be, is unacceptable in at least two distinct ways.
There is no reason to think that a few months or a year of filing experience will somehow permit a conclusion that the DOCX filing document can be treated as authoritative to the exclusion of a protective filing. It is discouraging to have to explain the same points repeatedly, but it seems necessary to do so.
First, DOCX is not a standard. You can take a DOCX file and open it in any of half a dozen word processors, and what will appear on the screen or be printed on paper will be half a dozen non-identical renderings. This means that it is a fool’s errand to try to use any particular DOCX file as an authoritative document for a US patent application. This situation is only worsened by the fact that USPTO’s way of rendering a DOCX file into anything human-readable (such as a PDF file) is by means of its own proprietary DOCX rendering engine, rather than by any open-source body of code. There is a level of distrust among users that USPTO might alter the internal function of its proprietary DOCX rendering engine at unpredictable times and without warning.
Second, there is no reason to think that some limited amount of filing experience between now and, say, January 1, 2023 could possibly fully explore the range of possible rendering errors that USPTO might make. There are many reasons for this. First, most the practitioners who file the most complex patent applications (those containing math formulas, chemical structures, Greek letters, tables, and other complex elements) have already resolved to avoid DOCX filing if at all possible for the simple reason of trying to avoid professional liability. There is no reason to think that the full range of rendering errors would be fully explored given that the handful of filers willing to gamble on USPTO’s filing system would be limited to those who have only the simplest of patent applications.
Third, even if some filing experience were to be accumulated between now and, say, January 1, 2023, that would provide no predictive power at all for things that might go wrong thereafter. Microsoft might choose to insert yet more proprietary functions into its implementation of DOCX after January 1, 2023, with unpredictable results as it relates to the USPTO’s proprietary rendering engine. Some other publisher of a word processor that is able to export DOCX formatted files might make their own implementation of some function in their own version of DOCX, again with what would almost surely be an unpredictable result as it relates to the USPTO’s proprietary rendering engine. Even if the publishers of word processors were to keep their code unchanged, what is there to block the USPTO from making hidden changes to its proprietary DOCX rendering engine? Any such events, occurring after January 1, 2023, would make a joke of any attempt to draw conclusions from the real or imagined success of a body of test filings prior to January 1, 2023.
From this discussion it is clear that the applicant needs to be given the opportunity to do a protective filing in a format that the applicant trusts. WIPO has provided this opportunity for over a decade under its PCT Administrative Instructions § 706, and the sky has not fallen. But the question that follows is, how long does the protective benefit of the protective filing need to last?
The protective benefit needs to last for the term of the patent plus the statute of limitations period. The June 2, 2021 notice contemplated that the protective filing might only be preserved for a year. From the point of view of applicants and practitioners this short period of time borders on the insulting. It strains credulity to suggest that the USPTO could offer any assurances that any and all errors introduced by the USPTO in its DOCX rendering would be detected by USPTO personnel within just a year of the filing of a patent application, or that that the applicant or practitioner could be certain of detecting any and all such errors within just a year. Examples of later times when errors might first become noticeable include the time of the 18-month publication, and the time of issuance of the patent itself. But from the point of view of professional liability, there is no choice but to recognize that many Office errors only get noticed at litigation time. This means that the ability to take corrective action with regard to a USPTO error in its DOCX rendering needs to be preserved until the risk of litigation has passed. This is at least the patent term plus the statute of limitations period.
Fee disincentives must be avoided. Given USPTO’s past behavior about DOCX, with a $400 penalty being proposed as the way to permit the USPTO to avoid an internal cost of a mere $3.15, it is natural to worry that USPTO might try to punish applicants or patent owners that try to show that USPTO made an error in its rendering of a DOCX file. We urge that any procedure for correcting an Office mistake in rendering or processing a DOCX file, relying upon a protective filing by the applicant, be free of any patent office fee.
Follow-on filings also need to be protected. Just as with the filing of new patent applications, if the USPTO decides to require that follow-on filings (subsequently filed documents) be filed in DOCX, then such filings also need to be protected. Thus, for any filing of a follow-on submission (subsequently filed document) where the document is being provided at USPTO’s request or requirement in character-base form (for example DOCX), the USPTO must provide an opportunity for the applicant to provide contemporaneously a protective version of the document in a user-selected format (for example PDF). The acknowledgment receipt for this filing must provide truthful message digests for both the actual uploaded character-based file and the actual uploaded protective file.
No more under-the-gun document reviews. One of the worst aspects of USPTO’s first DOCX procedure was that it forced the filer to review a USPTO-generated PDF file under enormous time pressure. This was unacceptable. Under USPTO’s third DOCX procedure, if a filer were somehow to realize that the USPTO is trying to slip into the file a modified DOCX file, the filer would only be able to avert harm by somehow reviewing that USPTO-generated DOCX file under enormous time pressure. That, too, is unacceptable. The USPTO must scrap its attempts to force the filer to agree, during the course of an e-filing session, to the supposed authoritativeness or finality of any document developed by the USPTO during the e-filing session.
In summary, we support USPTO’s goals of receiving character-based information from patent applicants and patent practitioners. We are discouraged that USPTO has failed to make use of many types of character-based information provided by filers in the past, for example with Issue Fee transmittals and Corrected ADSs. We are discouraged at USPTO’s serious missteps in all of its DOCX initiatives to date. We hope that USPTO can follow the suggestions set forth in this letter and we look forward to productive dialogue with the USPTO going forward.