2011 PCD Connectathon Lessons Learned

Ways to improve future PCD Connectathons

This is a collaborative page intended to capture every participant's insights on "lessons learned" from the 2011 PCD Connectathon. Please add your thoughts!

Increase realism of system performance under test

We need to "up the ante" on performing REALISTIC tests, and tests that are compatible with automated testing. In a realistic system:

• Consumers shouldn't need to manually add parameters one by one - they should be preloaded with the full Rosetta repertoire of parameters. This is most practical when the system has the capability of importing a file of parameter definitions (what the system needs to know to make an appropriate display based on, for example, the contents of OBX-3 Observation identifier and OBX-6 Units).

• No fair plugging in static contents where a real world system would be expected to generate appropriate values automatically. For example, NIST reports seeing some time stamps in messages that clearly were carried over from last year!

Improvements to the Connectathon Process

John Garguilo:

- Some vendors used preconfigured messages. This won’t work in the future when dynamic testing is implemented and messages will be real time (won’t be able to stop, configure, start). The PCC, in PIX/PDQ in Pre-Connectathon tests tools are currently available that go through a whole sequence and consequently do not allow this sort of.

- Some terms were missing in the tool and will be sent to Paul.

- MSH-3 ACK – test tool may be inconsistent with TF.

- PID-3.4, 3.5 need to clarify on whether required or optional; there is an inconsistency in the TF. This may need to reflect signing authority in the ADT message in contrast with generating the content.

- MSH-21 is not required.

- OIDs for PCD should become required and consistent with IEEE.

- HL7 in MSH-10 Control ID must be unique across devices. Within an ACK the received Control ID is provided in MSA-2 and the MSH is unique to the receiving system sending the ACK. Todd suggested that PCD can recommend adding some form of system ID to eliminate the need to also look at MSH-3 yet be unique. Al noted that MSH-10’s uniqueness requires MSH-3, so developers should use these rather than modify MSA-2.

- Testing should show that CT is indeed used on sending, receiving ends.

- NIST will explore providing logs directly into Gazelle and not require manual cut/paste.

Paul Schluter asked if message examples can be added to the TF. John Garguilo noted that he and Julien are beginning work on this. He noted that the Message Workbench can be used with pre-loaded messages.

Todd Cooper mentioned that there are some tricks to efficiently accessing log files. He asked if it is necessary to anonymize the messages that are posted, finding that vendor identities help prepare for testing. John Rhoads agreed, and indicated that messages inserted into the TF should be anonymous.

See also: Discussion in meeting PCD_TC_2011-03-09_Webex

Monroe suggested that Gazelle permit copying to Notepad for viewing, pasting.