Each year, US patent practitioners need to count up how many patents they obtained for clients in the previous year, for submission to the toteboards. What naturally comes to mind is the idea that maybe the USPTO’s Patent Public Search would permit a practitioner to get the answer. But USPTO’s Patent Public Search has, objectively, the worst search user interface that one could devise. Scroll to the end to see how you might use Patent Public Search to get your numbers for the toteboards.
In the old days (before 2021), it was easy to get numbers for the toteboards, using the USPTO’s Patent Full Text search system. PFT was familiar and user-friendly. It was also well documented. But in a cost-cutting move, the USPTO management decided to scuttle the Patent Full Text search system. This called for USPTO to provide some supposed counterpart to the Patent Full Text search system. The bright idea in USPTO management was to put some lipstick onto the system that USPTO’s patent examiners use, and force USPTO’s external customers to use that system. It is called Patent Public Search.
There are several reasons why Patent Public Search is a failure. To start with, it was specifically designed and laid out for the enormous computer screens that USPTO’s examiners use. USPTO’s examiners use two and even three computer monitors, each of which is approximately the size of a highway billboard. When USPTO management decided to force external customers to use this system, absolutely no effort was made to scale the panels of the user interface to function on any normal sized computer screen that any normal external customer would use. It means that the panels get scrunched to be small enough to fit on, say, a laptop computer, entire fields get forced out of view. It is quite literally impossible for a user on a normal computer screen to both (a) key in a search string and (b) view search results. If you make the search-results screen big enough to make use of it, you can only do this by shrinking the search-string screen so small that it is impossible to use, and vice versa.
Eventually the external-customer user works out as best they can some ways to cope with the failure of the Patent Public Search to scale intelligently from the bill-board-sized screen for which it was designed down to a normal-sized screen. The user gets accustomed to constantly needing to resize this panel or that panel a dozen times during a search session.
Which brings us to a second reason why Patent Public Search is a failure. The user documentation is inadequate. Whole categories of information that the user might need to know, to carry out even the simplest search that an external user might need to carry out, are missing from the user documentation. Over and over again the user encounters places in the documentation where a kind of search that an Examiner might need to do is fully and thoroughly documented. And over and over again what an external user finds is that whole categories of information are missing.
A few examples will illustrate this. Suppose, for example, that you are looking to search for plant patents whose issue dates fall within certain dates or for which the “attorney, agent or firm” contains the name of a particular law firm. If there were any place where USPTO’s documentation would explain how to search for a particular application type (such as “plant”) or a particular patent type (again such as “plant”), it would be on the Searchable Indexes web page, which if printed out on paper would span 17 paper pages. This enormous, run-on web page has a table that enumerates no fewer than 321 search fields that an Examiner might use. Over two hundred fifty of those search fields are directed to searches that only an Examiner would ever need to carry out.
In the old (good) search system called Patent Full Text search, the way to do the search was a field called “APT” for “application type”. With the old (good) search system, you could click to the place where search fields are documented, and you would learn that application type 1 is a utility patent, application type 2 is a design patent, and so on. The discussion for the “APT” field had a table listing the application types that the user might be searching for, and for each application type the numerical code was listed. It was in this table that you could cast your eye down a column, reach the word “plant”, and see that the numerical code was “4”.
If the developers of Patent Public Search had actually been tasked with designing a worthy successor for Patent Full Text search, then one of the steps of the design process (arguably the absolute most important step of the design process) would have been a line-by-line review of the user documentation for Patent Full Text. This documentation would have served as one of the “requirements” documents for the design of Patent Public Search. It would have been a requirement that each search field available to a USPTO customer in the legacy Patent Full Text would have a successor field in Patent Public Search. It would have been a requirement that the authors of the user documentation for Patent Public Search actually explain how to use this successor search field.
Returning to the “searchable indexes” web page, what one might have expected somewhere in the 17-page extent of the Searchable Indexes web page one might find the phrase “application type” and in close proximity, a simple table listing the application types and how to search for them. But no. The entirety of the documentation for the successor field (denoted “AT”) is:
Format: A.AT. AND glitter A1.AT. AND glitter
Comment: Searches Application Type – APT (1) in USOCR and Publication Kind Code – KD (1): in USPAT
Nowhere in this “format and comment” word salad do words such as “utility” or “design” or “plant” appear. I will reveal the actual answer and then we can return to this terse collection of words. The actual answer might be:
-
- to search for utility patents, use “b2.at.”
- to search for design patents, use “s.at.”
- to search for plant patents, use “p2.at.”
Returning to this “format and comment” word salad, the reader is left to speculate that “KD (1)” somehow means “go look in some obscure place to find out what to put before the ‘.at.’ depending on what you are looking for.
After some hunting around on some non-USPTO web sites, one stumbles upon a guess that “b2” probably does not actually get you all utility patents. Probably to get all utility patents, you have to do two searches, one for “b1” and another for “b2”, and combine the results, to get all utility patents.
Random clicking around reveals that the “KD” in this word salad might be a clue to look for another search field, namely the “KD” search field, which is said to mean “document kind code” for which the entirety of the format-and-comment word salad is:
Format: A1.KD. B1.KD. A.KD. Or I4.KD.
Comment: “A” for 2000 and earlier. B1 for patents without PGPUBs, B2 for patents with PGPUBs
Again, wholly missing from the word salad is any hint how this “KD” field in Patent Public Search might be used to search for, say, design patents or plant patents.
The hapless USPTO customer hoping to search for, say, plant patents might in desperation simply hit “control-F” to do a full-text search of the Searchable Indexes web page for the word “plant”. Astonishingly, the word “plant” appears only once on that web page. The place where the word “plant” appears is in a discussion of how to search for a document if you already know the patent number. In general, of course, you do not “already know the patent number” unless you are an Examiner working your way through an IDS that had been filed by an applicant. For a member of the public doing a search in Patent Public Search, the entire reason you are doing the search is that you do not already know the patent number. For a member of the public, the patent number is the “answer”, not the “question”. For a member of the public, the patent number is the “search result”, not the “search argument”. Oh, and in this sole place in the entire documentation where the word “plant” does appear, it says to use “PP” as the search string to look for plant patents. (As mentioned above, the correct answer is “p2”, not “PP”.)
In all of human history, any time that any system designer has forced users to migrate from a first search system to a second search system, one of the most basic requirements is to provide for the user a “search field mapping” listing each of the legacy search fields, and for each legacy search field, an indication of what the corresponding new search field. The USPTO developers failed to provide such a mapping.
The glaring omissions just discussed, along with dozens of other glaring omissions too numerous to list one by one, make it quite clear that nothing about the design and development process for what is now (misnamed as) Patent Public Search was ever meant for members of the public. Nothing about the design and development process for what is now (misnamed as) Patent Public Search was ever meant for users of the legacy Patent Full Text search system. It was never anything but an afterthought for what is now (misnamed as) Patent Public Search to be used by anyone other than patent Examiners.
Nowhere in the documentation is there even a hint or suggestion of how the public user would know whether to use “AT” on the one hand, or “KD” on the other hand, as a way to try to narrow down a search to, say, “design” or “plant” or “utility”. One gets the sense that the two search fields are very nearly interchangeable, except that they must not be identical in function given that they each have their own non-identical “format and comment” word salad.
As a third general example of the inadequacy of the user documentation for Patent Public Search, and as an example of the opaqueness of the underlying database search fields, consider the several non-identical search fields that each apparently somehow probe the “attorney, agent or firm” information that appears on the front page of an issued US patent. More or less at random the reader of the 17 pages of the Searchable Indexes web page stumbles upon no fewer than nine distinct search fields that might or might not serve this function. Here are the nine search fields, each quoted with some of its “format and comment” word salad.
-
- ATT – Attorney/Agent/Firm
- ATTY – Attorney Name
- FIRM – Legal Firm Name
- INAA – Legal Representative or Inventor
- LRAG – Legal Representative Name
- LRFM – Legal Firm Name
- LRFW – Principal Attorney Name
- LRNM – Legal Representative Name
- PATT – Principal Attorney Name
Nowhere in the Searchable Indexes web page is there any hint or suggestion how, if you are looking for, say, a “legal representative name”, you would know whether it is better to use LRAG on the one hand or LRNM on the other hand. Nowhere in the Searchable Indexes web page is there any hint or suggestion how, if you are looking for, say, a “principal attorney name”, you would know whether it is better to use LRFW on the one hand or PATT on the other hand.
One is left to speculate that the target audience for the system that is misnamed Patent Public Search includes not only USPTO Examiners but also perhaps the Office of Enrollment and Discipline. Some of the poorly document search fields include:
-
- LRCI – Legal Representative City
- LRST – Legal Representative State
I tried dozens of search texts in each of the search fields LRCI, LRST, LRAG, LRFM, LRFW, LRNM, and PATT, and every search came up empty. My best guess is that every one of these fields yields an empty search result unless the searcher is a logged-in USPTO employee.
As mentioned above, the Searchable Indexes web page has a table that enumerates no fewer than 321 search fields. Spot-checks of several dozen of these search fields lead to a speculation that all but a few dozen of the 321 search fields yield empty search results unless the searcher is a logged-in USPTO employee.
All of this communicates the unmistakable message that the participation by members of the public (customers of the USPTO) in this (misnamed) Patent Public Search system was never more than a mere afterthought, and that the only meaningful design effort or documentation effort by its developers was directed solely to USPTO employees.
Now we turn to how to get numbers for the toteboards. Suppose you are looking for design patents obtained for clients during 2024. Here are (redacted) search strings that various firms (here, “firmname”) have attempted:
( S.AT.) AND @PY=”2024″ AND firmname.ATTY.
(((“firmname”).firm. OR (“firmname”).inaa. OR (“firmname”).lrag. OR (“firmname”).lrnm. OR (“firmname”).lrfm. ) AND “2024”.py.) AND S.AT.
@pd>=”20231231″<=20250101 and firmname.firm. and S.AT.
The USPTO documention for “PY” suggests that this field could search for a “patent year”, but I tried some searches on “2024”.py. that yielded patents issuing in 2025!
The second query shows a firm’s search efforts to draw upon the search fields “inaa” and “lrag” and “lrnm” and “lrfm”, and as I say, my best guess is that any search on any of those four fields comes up empty for any public user.
Suppose you are looking for utility patents obtained for clients during 2024. Here are (redacted) search strings that various firms (here, “firmname”) have attempted:
(B1.AT. OR B2.AT.) AND @PY=”2024″ AND firmname.ATTY.
@pd>”20240101″ and @pd<“20251101” and firmname.att.
The former search seems to miss reissue patents, but perhaps the firm involved did not do any reissue work in 2024. The latter search assumes, I guess, that the firm’s practice does not include design patent work or plant patent work. Again one is left to speculate (given that the actual USPTO documentation fails to explain) what difference might flow from a search on “atty” or “att” or “firm”.
Finally, a search for plant patents might be along these lines:
firmname.att. and p2.kd. and @PY=”2024″