Now I have seen with my own eyes what others have reported. Until I saw it myself I could not be sure.
Yes, it really is true that Patentcenter sometimes lies about what exactly the document is that the filer uploaded to Patentcenter. The Acknowledgment Receipt sometimes lies.
This is a profoundly serious matter.
For what it’s worth, so far as I can see, the lying happens only with DOCX files. It looks to me like the lying does not happen with PDF files.
It seems that customers who use Patentcenter are going to need to scrutinize their ack receipts very closely between now and such time as the Patentcenter developers get this corrected.
With any government e-filing system, there is necessarily a relationship of trust between the filer and the government agency to which documents are being e-filed. When the filer uploads a document, the system provides an “ack receipt” (Acknowledgment Receipt) and the commitment from the agency to the filer is that no matter what happens later within the agency, including possible misplacing or corruption of the e-filed document, the ack receipt will permit a clear and undeniable record of exactly what the filer uploaded.
As of today, for me, this relationship of trust is lost with the USPTO. I uploaded a DOCX file, and the USPTO tampered with the DOCX file during the e-filing process, and when the USPTO generated the ack receipt, the USPTO system lied and said that what I had uploaded was the tampered-with DOCX file rather than the actual DOCX file that I actually uploaded.
In the discussion that follows I will provide the exact file that I uploaded, and I will provide the ack receipt, and I will provide the tampered-with file that is now in the USPTO system, namely the file that the USPTO system wrongly says is the file that I supposedly uploaded. I expect that this lying behavior of Patentcenter is completely reproducible and that anybody else who wishes to do so could reproduce this lying behavior of Patentcenter using these same files.
I was responding to an Office Action in a published US patent application. This means I can show you the actual files, because everything is a matter of public record.
The file that the USPTO system lied about is a file containing amended claims.
We start with the claims file that I actually uploaded. The name of the file is “P187089US00_claims_202203_draft.docx” and you can obtain that DOCX file here. If you extract a SHA512 hash of the file you will see that it is
You can open that DOCX file and you see that it has strikethroughs and underscores in it, compliant with 37 CFR § 1.121. I exported that DOCX file as a PDF and you can see the PDF file here.
After uploading the DOCX file, I clicked “submit” in Patentcenter, and I obtained the ack receipt. You can see the ack receipt here.
After receiving the ack receipt, I went into IFW to look at the claims in IFW. To my astonishment, the strikethroughs and underscores were gone! You can see it here. I took immediate action to deal with this problem, but if I had not, I must assume that what would have happened within the next few days is that some clerk at the USPTO would have scolded me by mailing out a Notice of Non-Compliant Amendment, for having failed to provide the strikethroughs and underscores.
A half an hour of forensic clicking and analysis reveals what happened. What happened is that the Patentcenter system tampered with my DOCX file and removed the strikethroughs and underscores. The Patentcenter system also tacked on a few extra characters at the end of the file name, namely “-APP.TEXT”. I have downloaded the tampered-with DOCX file and you can obtain it here. If you obtain a SHA512 hash of the file you will see that it is
At this point it makes sense to go back and look closely at the ack receipt. The ack receipt says this:
This Acknowledgement Receipt evidences receipt on the noted date by the USPTO of the indicated documents, characterized by the applicant, and including page counts, where applicable. It serves as evidence of receipt similar to a Post Card, as described in MPEP 503.
Yes, the language says that the USPTO received the “indicated documents” from me. The language says that the ack receipt is “evidence” that the USPTO received the indicated documents from me.
Except the problem is that it is a lie. The ack receipt says the USPTO received a document from me with a file name of “P187089US00_claims_202203_draft-APP.TEXT.docx”. This is flatly false. A true statement would have been that the document that the USPTO received from me had a file name of “P187089US00_claims_202203_draft.docx”.
Worse yet, the ack receipt says that the SHA512 hash of the document that the USPTO received from me was 00BC27FA516E8113334937863F3D8E868B6715B892DDB0750484103B09A416FE0C956EC2D7D60C6BBA7BE0B59D298599B476554CDA823E04F2E5D8F39DA3C8FD . This is flatly false. A true statement would be that the SHA512 hash of the document that the USPTO received from me was b95f48b4aaef60e5634f5c1a5b6c2be49b69b56b8ef3e54c3b866bed95f1c07ead202c874f423c9d1fbd14ad67ebc4419b638968f18c9ec7e830479aa137b945 .
As I say, others had reported such lying behavior in the USPTO systems with respect to DOCX files, and I wondered whether the reports were accurate or whether somehow there had been some misunderstandings about what had been uploaded or what was in the ack receipts. I kept thinking that I ought to follow this closely to see for myself whether such things could really be happening.
And here it is.
I have written this up as Patentcenter trouble ticket CP73 which you can see here.
This problem of Patentcenter lying about what DOCX file I uploaded brings to mind the USPTO’s ill-advised plans to impose a $400 penalty on any filer who files a patent application in any format other than DOCX. The penalty was scheduled to take effect on January 1, 2022 but got postponed to January 1, 2023.
Filers have been worried about this whole DOCX initiative at the USPTO. One of the worrisome things is that despite what the USPTO says, there is no actual “DOCX standard”. There is of course a mix of Microsoft Word standards for word processor formatting in a variety of versions of Microsoft Word. But depending on which word processor you happen to pick to display a randomly selected DOCX file (for example Google DOCS or Libre Office) what you will find is that the DOCX file may render differently on paper or on a computer screen. At the very least, it is very predictable that page breaks will happen in different places and that margins and formatting will vary. But what is more worrisome is that users report situations where chemical formulas and math equations don’t look the same depending on which word processor happens to have been used to view a particular DOCX file.
But what I saw today is a whole new category of scary. Yes, of course it is scary if I prepare a patent application in Libre Office and the USPTO’s rendering system uses Microsoft Word and the math equations don’t look the same. And yes it is scary if the USPTO e-filing system forces me to click “accept” on an adhesion contract in which I agree that whatever the USPTO changed in its rendering, I agree to the changes even if I did not notice the changes until after I clicked “accept”. Yes it is scary that the USPTO refuses to open-source its DOCX rendering system and instead keeps it proprietary (and might change it in the future without notice). Yes all those things are scary.
But this is a whole new category of scary. I can upload one DOCX file, and the USPTO system will tamper with the DOCX file, and the USPTO system will silently slip the SHA512 hash of the tampered file into the ack receipt, so as to make a false record that wrongly “evidences” that the tampered file is supposedly the file that the USPTO received from me.
There is a way to avoid all of this scariness, namely to pay the $400 penalty. Every time you file a patent application (starting January 1, 2023), you simply file it as a PDF and not a DOCX file. Yes you will have to pay the $400 penalty. But you will avoid having to click “accept” on an adhesion contract that agrees in advance to whatever there is about USPTO’s DOCX rendering engine that makes a math equation look different than the way it looked in your Libre Office word processor. And you will avoid having to worry about the USPTO system tampering with your DOCX file and slipping the SHA512 hash of the tampered-with file into the ack receipt to make it look like the tampered-with file is the one that you actually uploaded.
It is easy enough to describe how the designers of such a system could have been candid and truthful to the user. The designers of the system could have flashed a prominent warning onto the e-filing user interface. The warning might have said something like:
Warning! We have made substantial changes to your DOCX file. If you wish to proceed, you need to realize that we have discarded the DOCX file that you actually uploaded. It will not even be preserved in its original intact form in SCORE. What we plan to use is a highly modified DOCX file. The Acknowledgment Receipt will try to say that what you e-filed in the first place was the highly modified DOCX file with our changes. And no, we will not provide any change file or redlining to show what we changed. It is up to you to try to figure out what we changed.