Design Considerations

  • Error output from nodes is also in XML format and is managed by the Process Runner as a separate data stream.

  • Keep logic operators minimal to reduce user error. Looping and branching logic is implemented as predefined node behavior.
    • If products can react, this is implemented as an option in the reaction node, not as a loop in the Process Map.
    • Filter nodes implement branching constructs as output options. Substructure filter can have two output streams: matching & not-matching. Property filter can have output streams for less-than, greater-than, or a range.

  • Implement common usage patterns simply.
    • Multiple substructure filters can be "and-ed" together in a single node.
    • Property filter node will automatically get needed property calculations.

  • Data and error files are extractable archives containing data and processing history. These can be saved as file, stored in a database, or viewed through program interfaces.

Prev | Abstract | Next