Hahahaha. Please forgive me for my maniacal and semi-insane laughter…

Background: Two weeks ago I was teaching the PIUG Search Fundamentals Course along with John Zabilski in blessedly warm California and one of my talking point slides was about the US Patent Classification System (USPC). Officially, the system is dead and no longer being applied to patent applications being filed in the United States as of January 1, 2015.

However, almost nine million issued US Patents and another four million US Pre-Grant Publications have been classified within this system. It currently serves as a valuable search tool/method/heuristic/technique for US art that is distinct from every other method available to searchers, analysts, mappers, and others looking to review patent data. In addition, it is the only Classification system based on claimed subject matter rather than disclosed subject matter.

Today: All of these US documents have been classified within the Cooperative Patent Classification (CPC).

Tomorrow: The USPTO will disable the ability to search US patent documents by USPC in both their internal and external systems. These will be rolled out internally in Fiscal Year 2016 and externally in Fiscal Year 2017. (The external tools to be rolled out will replace the antiquated system currently hosted on the USPTO site.)

I was told that the only way to force US examiners to abandon the USPC system and fully switch over to the CPC system is to completely remove their access to the USPC. Wow. Thus the crazy, unwashed man living under the bridge laughter… Talk about a complete lack of trust in professionalism and training. It forces them to lose access to a very valuable tool that could still be useful, when used appropriately.


I am very glad that the commercial providers trust end users of the same data to make choices in a smart and reasonable manner. Options are important! See the discussion from last Fall when the subject was raised and several providers stated they will keep the USPC available as searchable fields. (Link)


  • No labels


  1. Very sad and an unfortunate decision by the USPTO that will undoubtedly  have an impact on the quality of patents granted by the patent office.


    On the flip side, that will be a value add that information professionals can bring to the table. It will fall on the "old timers" to teach the incoming generation about the USPC and it's advantages.

  2. Bizarre.  Both systems have their strengths and weaknesses.  Why not leave the existing USPC codes intact?  I would say that the patients are running the asylum since it's a bad analogy.  I agree that it's good that commercial providers will keep the archive intact.

  3. Thanks Dominic for posting this information.  I was under the impression from the various outreach efforts concerning the CPC that the US patent data would be reclassified according to the CPC all the way back to 1790.  Can any one shed some light on this?

    If indeed the entire US patent corpus will be reclassified, then presumably the CPC can be used to retrieve any US patent document.  But, I do agree with both Dominic and Andrea that the USPCs, even only as legacy data now, still have their place in helping to find documents of interest. Not only because of the different approach taken in assigning US classes when compared to the approach taken with other patent classification systems, but also because of the manner in which reclassification occurs.  I seem to recall that the reclassification of the US patent data would rely on imperfect concordances and algorithms rather than human beings reviewing the patent specifications and applying the most relevant CPC codes.  That human intellectual input, particularly for those US patents granted long before any of the other patent classification systems were even created, would be lost if USPCs could no longer be searched.  Fortunately, some of the database providers have indicated their committment to ensure USPCs can be searched well into the future.

    If the entire US patent corpus is not going to be reclassified according to the CPC, then the decision to abolish access to the USPC within the USPTO's internal and external resources is a drastic measure to take as the US Patent Examiners will need to consult third party commercial tools that retain USPC searching capability.  This will add time and extra effort to the Examiner's day and it would not be surprising if some were tempted to simply rely on keyword searching of the pre-CPC prior art. This would not help the USPTO meet the aims of its recently announced Enhanced Patent Quality Initiative program.  It would also seem a strange outcome that the very organization examining the current patent applications would not be in a position to search the prior art it itself had published prior to Jan. 1, 2013 by either the CPC or the USPC.

    So what's the deal?  Will the US patent corpus be reclassified or not?

    1. Susanne,

      All of the US art has been classified in the CPC. All of us here in the US are “drinking the Kool Aid” and the CPC should now be the primary classification tool used to search US art.

      This is difficult when you have 10-50 years of experience using a different (USPC) system. Many of the subject matter experts at the USPTO (and quite a few external experts as well) fall into this category and these are the likely targets of this field code elimination plan. I fear that some of these true wizards of their arts will just retire rather than learn new tricks and the youngsters will have one less opportunity for mentoring.


      You do bring up a major issue with the switch, but one we must live with… A large amount of US art classification in the CPC was done using statistical mapping. This is one reason the old, legacy USPC system is still important to remember.

      For example, when an art area in the USPC was based on structure and the new CPC is based on function, the mapping resulted in low accuracy as the two systems do not correspond to one another. This is particularly true for pre-1920 art which was 100% classified in the CPC system using statistical mapping.

      In this example, for a high precision search of US art based on function, you should use the CPC. However, for a high precision search of US art based on structure, you will still want to use the USPC. This is the choice and functionality that will be removed from the US examiners “adaptive toolbox”.

      What our mutual friend Stephen Adams wrote fourteen years ago in World Patent Information about the US and IPC systems remains true today with regard to the US and CPC systems. “The US and IPC classifications do things differently, and the skilled information scientist should spend some time getting to know both systems in their particular technical field.” However, this need will slowly fade as the USPC system stops being applied to new cases moving forward.

      (To be honest, we are currently in a “golden age of classification” for US art with three systems covering all US art. No “golden age” lasts forever.)

      See you in May,

  4. Follow Up: Several good people at the USPTO have approached me in the past week and told me that they are doing their best to make sure US patents are searchable using the US Patent Classification (USPC) system both internally and externally in the future. They understand that even though the system has died, it still provides useful information from the grave...

  5. USPC, rest in peace :'(

    What is life like under CPC?

    Source for UPC-CPC mapping: Questel Orbit.


  6. Tribute for the United States Patent Classification (19001 - 2015)

    Even though your boss at the USPTO threw you under the bus, aliens and non-aliens around the world within the patent information community thank-you for being highly relevant in our patent analytics work. You shall be fondly missed and remembered. Rest in Peace.

    From Rex Yeap, Singapore.

    1 O’Hara, F. J. (1977). Views and over-views on/of U.S. government documents. Government Publications Review (1973), 4(3), 263–270. http://doi.org/10.1016/0093-061X(77)90010-7