• Added more tuning and stability-related options to the scanner instructions for the Reader (see details below.)
• Improved performance by submitting several small documents to the Reader at the same time. Prior versions of Postmaster code would submit OCR documents to an engine serially (or one at a time) regardless of the system resources required to process a document. This version of code will prorate the points consumed by any job based on the number of "chunks" in the document and may process several smaller documents at once.
• This version of code does not require the basic image instructions to be copied for every scanner; only entries that override the default values need be created for scanners. BuildImageInstructionsForAllScanners.sql is no longer required for this version.
• Upgraded OCR engine for general performance and fix a problem with some text on pages being skipped during OCR.
• Restructured Reader to fix problem of unexpected page removal when barcodes are redacted.
• Added thread timeout options to help prevent hung processes from blocking further processing.
• Fixed a problem in the Queue Builder that would stop processing if a machine hosting one of the scan folders was shutdown or unavailable. This version will continue processing with any available scanners.
• Resolved an issue with network socket handling that could cause process instability including retries and documents being assigned a HOLD status.
• Improved engine installer so previous versions of MM3 can be updated without first uninstalling.
• The ImageInstrType.ArgumentAllowedValues for --ConcurrentChunks should be 1,2,3,4,5,6,7,8,9,10.
• The ImageInstrType.ArgumentNotes for --ConcurrentChunks should read "The number of chunks that may be processed simultaneously on the smallest engine, typically 1/2 to 1 times the number of processor cores."
• The ImageInstrType.ArgumentNotes for --AllowSimpleCopy should read "True allows fastest handling of stacks but could cause processing failure if the input file is poorly formatted. False causes the separator to create a new file to eliminate formatting errors."
• The ImageInstrType.ArgumentNotes for –MaxSeparatorThreads should read "Controls multi-threading behavior of the Separator, typically 1/2 to 1 times the number of processor cores."
• The ImageInstrType.ArgumentNotes for --UserMargin should read "Specifies the page margins (in pixels) used during blank page detection. Change this value only under advice to address blank page detection problems."
• This version may redact Postmark printed barcodes when a printed document is scanned. This may change in a future version.