Background

For sites that need to split processing between multiple installs of SOCR, the preferred method is to split by cabinet or folder (i.e. have one set of folders that one install of SOCR is responsible for, and another set of folders that another install is responsible for).  This isn't always possible, and another strategy is to use date filters to segment the processing.  The idea here is to allow configuration of a filter that completely blocks the instance of SOCR from finding the documents that it shouldn't process (they won't appear in the second Symphony OCR database at all).

Note  DMSes that force the modified date to change (i.e. NetDocuments) will ultimately end up with the document being discovered by the other SOCR install.  So the document will get analyzed a second time and moved immediately to the Processing list.  Not the end of the world, but it'll involve unnecessary downloading of the file.

Resolution

Currently, this functionality cannot be configured through the UI.  In order to configure this option

• Update Symphony OCR to version 6.4.81 or higher (if necessary) on both workstations that will run Symphony OCR
• Go to the license page and click "Save Settings" (this updates the settings.xml file to include the necessary entries)
• Close Symphony OCR
• Navigate to the C:\Program Files\SymphonyOCR\Config
• Open the settings.xml in notepad
• Look for this entry:
  <finderHandler cutoffTimeHigh="NOCUTOFF" cutoffTimeLow="NOCUTOFF"/>
  
  you can change either or both of the NOCUTOFF values to be a date.  The date is specified in GMT timezone.
  
  For example:
  
        <finderHandler cutoffTimeHigh="12/31/2010" cutoffTimeLow="NOCUTOFF"/>
  
  will make it so this instance of SOCR will only see documents with modified dates prior to 12/31/2010.
  
  You'd then configure the second instance of Sympony OCR like this:
  
        <finderHandler cutoffTimeHigh="NOCUTOFF" cutoffTimeLow="12/31/2010"/>
  
  

This will ensure that one instance of the software is only processing documents with a later modified date than the one specified, and the other instance is only processing documents with an earlier date than the one specified.