What we wrote to one of our instructing foreign patent firms today about DOCX

Two days ago I published (blog article) the text of a letter that our firm sent to some of our instructing foreign patent firms.  One of the firms wrote back to us about this, asking whether the “auxiliary PDF” option in the DOCX filing path might alleviate the DOCX risks sufficiently to permit following the DOCX filing path, and thus to permit avoidance of the $400 non-DOCX penalty.  Here is how we responded to the instructing foreign patent firm: 

Yes, if a foreign instructing patent firm were to instruct their US counsel to do the filing in DOCX, and if US counsel were to agree to carry out the instructions, I think it would be a Best Practice for the US counsel to include the aux-PDF in the filing package.  Indeed it would be extremely foolish for US counsel to fail to include the aux-PDF in the filing package.

It is extra work for the US counsel to do the filing using DOCX and the aux-PDF, compared with the amount of work required of US counsel for the trusted legacy PDF filing path.  For the DOCX filing path, the US counsel uploads a DOCX file which we call D1.  But that does not become the patent application at the USPTO.  What happens during the e-filing process is that the USPTO modifies that file and creates a new DOCX file which we call D2.  If the US counsel were to click “submit”, then the D2 file would become the patent application at the USPTO.

This means that before clicking “submit”, US counsel must carry out a line-by-line comparison of the D2 file with the D1 file, during the e-filing process.  This takes time and costs money.   Probably this task must be carried out by an attorney, and cannot be delegated to a paralegal.

During the DOCX e-filing process, the USPTO system generates a “Feedback Document”.  This Feedback Document contains comments from the USPTO system about things that it did not like about the D1 file.  But the USPTO does not promise that the Feedback Document will disclose all of the things that the USPTO changed when it created the D2 file.  So the danger of the D2 file differing in some important way from the D1 file is not eliminated by closely reading the Feedback Document.

Some US counsels might feel that during the e-filing process, it is necessary to study and review the Feedback Document.  if this study and review is carried out, this takes time and costs money.  The US counsel would probably judge that this task must be carried out by an attorney, and cannot be delegated to a paralegal.

You mentioned the aux-PDF.  Yes, the Best Practice for the DOCX e-filing process includes uploading  an aux-PDF file.  The problem of course is that probably the D1 file is not exactly the same as the DOCX file that foreign counsel sent to the US counsel.  Probably the US counsel would find the need to make some changes to the DOCX file.  Maybe changing from A4 to 8½x11.  Maybe changing the line spacing.  Maybe inserting section headings.  Maybe correcting a grammar error in the document.  Maybe editing some claim language.  Maybe eliminating multiple dependency in the claims.  Maybe inserting incorporation-by-reference language from an earlier patent filing.  This means that the DOCX file will have changed since it was received from foreign counsel.  The DOCX file that US counsel uploads to Patent Center (which we call D1) will thus not be identical to the DOCX file that was received from foreign counsel.

One of the USPTO rules is that the aux-PDF file must exactly match the D1 file.  If the two files do not exactly match, then probably the USPTO will refuse to give value or meaning to the aux-PDF file.   This means that US counsel need to add a second pair of eyes to the DOCX e-filing process, to track the version changes of the DOCX files and to track the corresponding version changes in the supposedly matching aux-PDF files.  It is extremely important to make absolutely sure that the version of the aux-PDF file that gets uploaded is exactly the same as the version of the DOCX file that gets uploaded (which we call the D1 file).  This takes time and costs money.

On the other hand if the USPTO e-filing path is the PDF path that we have been using for the past 20 years, where a PDF file is the patent application, then none of these extra steps is needed in the e-filing process.

If, for example, a US firm had historically charged a flat fee to do the traditional PDF filing, then to cover the extra work for DOCX and aux-PDF, the US firm would surely need to charge a higher flat fee.  The higher flat fee for the DOCX-and-aux-PDF path would surely be at least $400 more more than the historical flat fee for the legacy PDF filing.

But the extra work is not done when the application is filed.  When the 18-month publication of the US application happens, it is necessary to proofread the publication line by line to check to see whether the USPTO rendering robot made mistakes in rendering the D2 file into the publication.  The US counsel would need to be paid to do that work.

And the extra work is not done when the proofreading of the 18-month publication has been done.  When the USPTO issues the patent, it is necessary to proofread the issued patent line by line to check to see whether the USPTO rendering robot made mistakes in rendering the D2 file.  The US counsel would need to be paid to do that work.

It might be thought that a thorough proofreading of the publication would permit everyone to relax and to skip the thorough proofreading of the issued patent.  This is not so.  The so-called “black box” rendering engine that the USPTO uses for robot processing during the filing process, and that the USPTO uses for robot rendering of the publication, and that the USPTO uses for robot rendering of the issued patent, is in a state of constant change.  As of December of 2023, according to the blog of USPTO Director Kathi Vidal, the black box rendering engine had reached version 18.  At least one version change occurred in January of 2024.  The USPTO does not log the version changes of its black box rendering engine and does not reveal what changes are made in each update.  It is to be expected that the rendering engine will have had (undisclosed) version changes between the date that the robot renders the publication and the date that the robot renders the issued patent.  For this reason, the line-by-line proofreading of the issued patent is needed even if a line-by-line proofreading of the publication had already been carried out.

Any of a number of events might lead to a patent owner needing to get the USPTO to correct a mistake in an issued patent.  The events might include:

• • noticing a USPTO mistake as a result of a line-by-line proofreading of the publication shortly after the publication;
  • noticing a USPTO mistake as a result of a line-by-line proofreading of the issued patent shortly after issuance;
  • noticing a USPTO mistake as a result of licensing activity;  and
  • noticing a USPTO mistake at litigation time.

When a USPTO mistake is noticed, how does the patent owner get the USPTO mistake corrected?

If the filing path was the legacy PDF filing path, then the correction is simple and the USPTO is not allowed to say “no” to the correction.  The patent owner compares the PDF patent application with the issued patent, and if they do not match, then the applicant is entitled to a Certificate of Correction from the USPTO.  The USPTO is not permitted to fight this.  The USPTO must simply issue the Certificate of Correction.

If the filing path was DOCX without an aux-PDF, and if more than one year has passed since the filing date of the patent application, then no correction is available.  (It will be appreciated that usually litigation time comes more than a year after the filing date of the patent application.)

If the filing path was DOCX with an aux-PDF, then the path to seek correction is filing a petition.  The USPTO might grant the petition or the USPTO might dismiss the petition.  If the aux-PDF does not match the D1 file, then the USPTO might use this as an excuse to dismiss the petition.  The USPTO might find some other reason to dismiss the petition.

In the e-filing process, the USPTO discards the D1 file.  The D1 file is not preserved in USPTO’s systems.  The only DOCX file that the USPTO preserves in its systems is the D2 file.  To be able to prove that the aux-PDF matches the D1 file (not the D2 file), the US counsel need to preserve the D1 file locally, in perpetuity.

In the e-filing process, the USPTO discards the actual aux-PDF file that the US counsel uploaded.  The actual aux-PDF file that the US counsel uploaded is not preserved in USPTO’s systems.  The file that appears in Patent Center, that purports to be the aux-PDF file, is not the same as the actual aux-PDF file that US counsel uploaded.  The purported aux-PDF file is a different file size and has a different file name than the true aux-PDF file that the US counsel actually uploaded.  The purported aux-PDF file has been resampled and halftoned.  So it probably looks different from the actual aux-PDF that the US counsel actually uploaded.  To be able to prove that the aux-PDF matches the D1 file, the US counsel need to preserve not only the D1 file locally but also the original aux-PDF file locally, in perpetuity.  US counsel would need to be paid some money to do this local file preservation of these two files.

There is the risk, as I mentioned earlier, that the USPTO might dismiss the petition for correction.  The US counsel will need extra malpractice insurance to cover this risk, and to cover several more risks:

• • the risk that US counsel could be accused of having failed to do a smart enough comparison of the D1 and D2 DOCX files during the e-filing process;
  • the risk that US counsel could be accused of having failed to study and review the Feedback Document sufficiently during the e-filing process;
  • the risk that US counsel could be accused of having uploaded an aux-PDF file that fails to match the D1 file;
  • the risk that US counsel could be accused of having failed to do a smart enough proofreading of the 18-month publication;
  • the risk that US counsel could be accused of having failed to carry out the proofreading of the 18-month publication;
  • the risk that US counsel could be accused of having failed to do a smart enough proofreading of the issued patent;
  • the risk that US counsel could be accused of having failed to carry out the proofreading of the issued patent;
  • the risk that US counsel could be accused of failing to do a good enough job of informing the client of the risks associated with the risky DOCX filing path instead of the trusted legacy PDF filing path.

For US counsel to be willing to take all of the professional liability risks that are described here for the DOCX filing path, probably two things would need to happen.

First, it would be necessary to prepare a consent agreement in which the client would be told about all of these risks.  In the agreement, the various risks would be described.  The client would be agreeing to take all of these risks and would be agreeing not to hold the US counsel responsible for all of these risks.  The client would be advised to consult separate counsel for advice about these risks.  In the document, the client would give consent to proceeding in the DOCX path despite the risks.

Second, the US counsel would need to be paid a higher professional fee for the DOCX filing path, as compared to the professional fee for the legacy PDF filing path.  The higher professional fee would need to reflect the extra work required for the DOCX filing path.  But the higher professional fee would also need to reflect the substantial professional liability risks for US counsel that flow from the DOCX filing path.

I think there are many US counsel who will simply decline to do the DOCX filing path even with a signed consent agreement and with a higher professional fee.