Classification Metadata?

Dominic,

There are at least two possible approaches to this:

a) if the revision of the class in question is the first and only one since the original search date, you can use a source like STN's US full text file, which stores both the original class and the current class.  But as you have observed, this loses any intermediate steps, so,

b) it sounds like a case for tracking through the Classification Orders to see whether there have been any Orders issued since the date of the earlier search, which impacted upon the class(es) which were used in that search.  If you can identify the Orders concerned, these will tell you where the patents were re-classified to under that change.  Then you need to repeat the process to see if any of the classes created by the first Order have themselves been the subject of any later Orders, resulting in further distribution of the prior art. Keep repeating the process until you either have a headache or reach the current class structure.

Seriously, this illustrates why I have been pressing for ages that all the classification schemes should maintain a complete audit trail in order to roll back the identify of current marks to what they would have been on a specific point in the past.  The IPC is some help with the revision concordance lists, but these are not so easy to track since the 2006 reform.  ECLA has no history of past class structure, and neither I suspect does FI.  It came out in conversation that the Canadian Patent Office may be one of the few offices which is maintaining a full back-file of all the IPC classifications appended to any given document since publication, but it was not clear whether they would be prepared to offer this as a service to outside parties.

Hope this helps

Stephen

Dominic,

Contact Varona Stevens in the Editorial Division. SIRA-DCD Varona.Stevens@uspto.gov. regarding the Classification Order Archival Report (Abolished and Established Classes and Subclasses for Classification Orders 1 through 1895). Alternatively, Office of Patent Classification Divisions, Data Control Division may be able to provide a better suggestion.

It seems that a search of the Classification "Numeric Class" for the year and "Order No.", comparing the result with the Classification Order Archival Report, then a crosscheck with the Reclassification Alert Report would provide the detail you might need.  At the least, it might narrow the date range of any classification changes to a two month time frame.

http://www.uspto.gov/patents/resources/classification/Classification_Conatact_Numbers.jsp

Best of luck, Sherri Voebel

Thank you to the responders (both here and through direct communications).

Unfortunately it is apparent that nobody (except perhaps Canada) has taken it upon themselves to differentiate their classification data from their peers... (Hint, hint.)

I've survived my situation, but I would ask someone to please create a way to easily avoid the following specific situation which I encountered because it will happen again with the pace of reclassification occurring in the various patent offices and potential merging of the US and European systems in the future:

1)  I was performing a US Freedom to Operate (FTO) search on a method for electrically testing chemicals while they are in a container.  I manually reviewed the claims of all 3000+ US patents and applications in US Class 422, Subclass 102 (Link) on January 10, 2011.

2)  On January 11, 2011 this subclass was abolished by the USPTO and the 3000+ references previously reviewed are now intermingled with the additional 7000+ references from Subclasses 99, 101, 103, and 104 as detailed in the Classification Order (Link) of November 2, 2010.

3) I now have no way of identifying the references previously reviewed...

*** This subclass abolishment and the new classification scheme had not propagated through all the different vendor databases and thus, I was able to identify the reference numbers.  But if the time gap between the initial searching and later searching was larger, this method would not have been available.