Patent applicants and practitioners continue to resent the USPTO’s $400 penalty that it plans to impose upon any patent applicant that has the temerity to try to establish a PDF file as the “controlling” document for the patent application being filed. The penalty was going to start a few days ago, on January 1, 2023. The USPTO blinked and has postponed the penalty to a starting date of April 3, 2023. As of right now, that is when this penalty will begin.
The other day I realized that there is a pretty strong analogy to make here between this $400 penalty and another bane of daily life — car rental companies gouging renters with the “collision damage waiver”.
It is of course impossible to predict whether a particular car rental will result in damage to the rental car. Most rentals start and finish without anyone becoming aware that anything bad happened.
Likewise it is impossible to predict whether version 19 of USPTO’s proprietary PDF rendering engine, which USPTO will place into service without advance warning at some unknown future time, or some existing defect in the engine (which Director Vidal recently revealed is at version 18), will lead to a practitioner getting sued ten years from now in a case where the practitioner failed to pay the extra money to be able to file something that the practitioner trusts to be the controlling patent application. USPTO says, and it is probably true, that in the majority of issued patents going forward where the practitioner gives up and files in a way to avoid the $400 penalty, the practitioner will not actually get burned, either because the USPTO proprietary engine won’t screw up a particular case materially, or the screwup won’t ever get noticed.
In the case of car rentals, the customer can pay the “collision damage waiver” to the rental company. This makes the customer’s financial risk go away. This is an astonishingly profitable add-on from the rental company’s point of view. Observers have suggested that at least 90% of the fee is pure profit to the car rental company.
In the case of the USPTO with DOCX, the practitioner can pay the $400 fee which makes it possible to change the “controlling document” from USPTO’s DOCX file to the practitioner’s trusted document (for me, either ODF or PDF). This restores the practitioner’s risk back to what it was before this whole DOCX abomination.
The $400 fee is likewise astonishingly profitable from the USPTO’s point of view. USPTO admits its OCR cost for a filer who files a PDF patent application is $3.15, so the other $397 is pure profit.
Most readers are probably quite familiar with USPTO’s DOCX abomination. The idea is that starting on April 3, the filer will be expected to refrain from filing any document that the filer actually trusts (typically a PDF file) and will be asked to file only a Microsoft Word file in DOCX format. The USPTO will then render the DOCX file into PDF using a proprietary PDF rendering engine and DOCX validation engine that has by now changed 17 times and will surely change without notice at random times in the future. The USPTO has in mind that the “controlling” document will actually be whatever PDF file (or “validated” and modified DOCX file) the USPTO later generates from that originally filed DOCX file.
This puts patent practitioners at grave risk of things blowing up later and getting sued for malpractice. Version 19 or 20 of USPTO’s rendering and validating engine might change a square root symbol into a smiley face, and if it did, maybe the first time it would get noticed is TYFNIL.
As things now stand, I believe that most patent practitioners will adopt a routine practice of paying the $400 and filing in PDF. This eliminates the risk of unpredictable future USPTO behavior changing a square root symbol into a smiley face. It is like always paying the gouge-priced collision damage waiver to the car rental company.
The obvious solution to this problem is that the USPTO should save the docx as filed. I thought I read somewhere they weren’t going to do that but the USPTO should. Docx files are not a heavy data burden.
Yes of course the USPTO should preserve (through the life of any issued patent plus a statute-of-limitations period) whatever computer files the applicant provided to the USPTO on filing day. Astonishingly, the USPTO most recently proposed that after a year, it would discard every computer file that the applicant filed except the particular computer file that it says “controls”.
But preserving the word processor file that the filer filed will not protect the practitioner fully if the format is a file ending in “docx”. The problem is that there is no standard for “docx”. Three different people could open that file later and see three different things on the screen. Page breaks and line breaks will be in different places. Other things are likely to be non-identical in the renderings to human-visible from “docx”.
If you were to set a goal of having the “controlling” document be a word processor file, it is a fool’s errand to try to do this with “docx”.
There is a present-day word processor format that is the subject of standards-setting activity, and that can be rendered unambiguously into human-readable form now and in the future. But that format is not “docx”. It is OpenDocument Format. Microsoft Word can export ODF files. All of the other present-day word processors either use ODF as their native storage format or can export ODF files. If USPTO were to set a goal of having a word processor file be the “controlling” document, then there is no honest present-day option but ODF.
Thank you. I saw your other post about the variability in docx files after I sent in my post. I assumed that MS Word is the de facto docx tool, and wasn’t aware of the variability but your arguments make perfect sense. So this means the USPTO needs to also preserve a user supplied pdf file.
I’m fundamentally OK with filing in docx because it should provide unambiguous text data, so I would like so see this work somehow.
One concern is “the variability in docx files” given that there is no actual standard for that word processor format (it is a proprietary Microsoft thing). But there is the distinct thing that USPTO’s present approach is that they render the applicant-provide DOCX file into “something else”, and it is that “something else” that USPTO says will “control”. The rendering into “something else” is carried out by USPTO’s proprietary engine. Which is by now in version 18 according to Director Vidal. Meaning it has had 18 different behaviors at various dates in the past, and will have still more non-identical behaviors in the future.
You write that docx “should provide unambiguous text data”. I am not clear on your meaning here. Do you mean “should” in the sense of “it doesn’t do it now but that’s not the way it ought to be”? If that is what you mean, then what would need to happen is Microsoft surrendering control of the format to a neutral standards-setting body. I am not aware of anybody who thinks Microsoft will ever do that.