ICAR Animal Data Exchange Working Group – Technical meeting 28 February 2020

• Apologies: Andreas, Henning, Beate, Kor, Sven
• In attendance: Andrew, Anton, Arjan, Bert, Conny, Craig, David, Erwin, John, Marvin, Thomas.

1. Lactation (GitHub issue 65)

• Erwin had edited the issue and added notes for decisions.
• Bert had supplied information from previous work
• Added the icarMilkRecording method with various attributes

1. milkRecordingProtocol
2. milkRecordingScheme
3. milkingsPerDay
4. milkSamplingScheme
5. samplingMoment
6. recordingInterval
7. certified
8. milkingType

• The group decided that as enough work had been done previously to represent ICAR approved recording schemes, these fields should be kept, although optional.
• Change recording interval to an integer number of days rather than a long enumeration.
• Discussed weighted averages of the test days - e.g. urea average, somatic cell count average. Decided this could be left out at the moment.
• lactation result code (e.g. 100 days, 200 days, 305 days) - we decided this should be an enumerated attribute.
• parity or lactation number - discussed this is not (necessarily) the parity of the animal, but could be lactation parity.
• Noted that all results here are calculated/aggregated for a lactation - there are separate events for milking visit and test day result.
• Erwin will implement (PR #82)

2. Documentation (GitHub issue 72)

• It would be helpful to describe what messages there are and how they relate to each other.
• Marvin described how an overview would be helpful for new implementers.
• Structure - category, messages and their intents. Overlap with the URL scheme internal documentation.
• Discussed how to automate documentation from GitHub actions. Decided that might be too hopeful at the moment.
• Andrew to start drafting overview, Marvin to review.
• Add instructions for how to use issues and participate
• --> Andrew to ask wider ADE group at next meeting what they think is missing.

3. Restructuring folders and type/resources (GitHub issue 66)

• Remove old JoinData and DataLinker inspiration folders.
• We liked the suggestions in this issue. Anton will undertake. (now done by Anton, PR #79)
• Discussed how implementations already developed by CRV and JoinData that use the messages but not the collections could be updated. Pagination is optional (JoinData does not currently support this), so this does not need to be a blocker to implementation. JoinData can merge information from multiple sources, in which case pagination can become difficult.

4. Versioning Scheme (GitHub issue 17)

• When Andrew documented the group's previous decisions regarding versioning, he suggested making the major version something recogniseable such as "ADE-1" rather than "V1". The group agreed this approach.

Agenda for next meeting: release of version 1 (ADE-1), workshops, then suggestions for future work.

*Next meeting delayed until 20 March at 9am CET