Originally posted by: SystemAdmin
To the best of my knowledge, at this point there is no plan to change the way that HCIP type 5 checking is defined. As always, if there is a strong need for change, Product Management should be consulted.
That said, there are technical reason for the changes to be made in the trees - and it is partially a function of the version of HCIP that you are running.
HCIP versions prior to 4.0 (HCIP2.1, HCIP3.0.x and HCIP3.1.1) used Compliance check validation out-of-the-box as a stand-alone validation tool. With HCIP4.0 and beyond, efficiency improvements were made by adding a pass of type tree validation as a prequil to Compliance check. This helps speed processing becasue tree based validation is (relatively) much faster than full layered Compliance Check Validation. So if you throw the data against the type tree and only pass data that fails validation to Compliance Check for reporting, things are speeded up.
Why this matters: With HCIP4.0 and beyond it is very important that your data type trees (such as hipaa_x12_4010a1_type_4.mtt) and your compliance check components (such as hipaa_x12_4010_segment.mtt and the qualifier rules) stay in sync. This is an important consideration in the design of the type 5 approach as the HCIP developers could not rely on a simple table based approach that added rules to the qualifier table only (used in Compliance Check) because the new rules would not be useful in the data type tree (such as hipaa_x12_4010a1_type4.mtt). They needed to use a tree based approach to handle both data type tree changes and Compliance Check changes. By setting up a simple recommedation (and example) for updating both processes type trees - this was achieved.
The nature of the type 5 validation:
- Multiple lists from multiple data sources
- Frequent changes to lists
- Not all customers want to implement all lists
...Make it prohibitive for Ascential to maintain a common set of trees with type 5 checks and distribute in a reasonable manner (IMHO)
The recommendation was therefore to have customers add and update the lists they want on the schedule they want. That said, if I were a customer I may want to consider adding a process (as I believe was suggested in this thread) around the import of these lists that would include: loading this data into a database (with version, source, and tracking into), and creating the table to be imported into the type trees from a query of the more robust database.
I do not know if anyone has officially asked Product Management to embellish their type 5 example to do this (as an Enhancement Request)... but it never hurts to ask.
#IBMSterlingTransformationExtender#IBM-Websphere-Transformation-Extender#DataExchange