2019-07-26 - System Operations and Management SIG Agenda and Notes

• new members ?
• note taker: Phil
• Review Action Items ? Not today

Tod and me brought our issue into the Implementers Group on Tuesday.

They said they will make additions to the Spreadsheet (FOLIO Source Data and Integrations Spreadsheet ) until July 31st.

The spreadsheet is considered "high level" and we are supposed to do the low-level, viz. technical, analysis.

I transmitted all information about integrations and JIRA-Tickets from that spreadsheet to our analysis wiki site: Integrations - Open features and known gaps

I had no time to further look at it, at the moment it's just 1:1 what was in the spreadsheet. I hope I can make some first re-arrangements and see where are gaps and what needs to be prioritized (if I can judge that).  Otherwise I hope we can do this work tomorrow in the group meeting.

Uschi made a comment about RFID integration which we should include.

Meeting Notes

We have to test these things out.

What is under development, what is not → GAP analysis

Brandon: Wouldn't prioritizing these integrations fall under another SIG ?

In some cases yes, though implementers would not have the technical background to provide good requirements / estimates.

Tod: Do we have the pieces on the development channel to allow it to happen ? We have to cut out unpleasant surprises. Prioritizing is not the point. We have to decide what is missing.

Jo: We need to integrate data that all libraries need. It's an "information gateway"

Ian/Phil: Some tools could open up doors. We need to give architectural guidance.

Tod/Phil: We have to consider lead times / trajectories. We have to keep in mind the MVP concept.

We should think of these from the FOLIO product perspective, not just our institutions.

• Exports of data files will be a MVP go-live requirement (so, ready by July 2020 for several institutions).
  • Are we talking about pushes or a pull? Mostly pushes for now, like an ETL. Could it be a use case for the Workflow module? It opens the door for a lot of kinds of data integrations.
  • Many schools have Banner, and a Banner <> Voyager file push.
  • At the moment, we can talk to Okapi.
• Can we get a filtered list of JSON user objects? Yes, CQL query from the Users app. 
• One barrier is that FOLIO specs are still a bit of a moving target
• Is the SysOps SIG the right place to do deep dives on financial integrations?
• Maybe as a partner on estimating the technical groundwork needed.
• Are there Edge APIs we should be developing? Such as a patron data export.
• Exporting files (such as to financial systems) - have we figured out a common way to do that?

• There’s a proposal to fund a prototype to enhance the Workflow tool along these lines (6-8 month lead time needed, though)
• Maybe we could approach it as MVP basic file generation to make the July 2020 go-lives
• EDI is an important need that could be difficult - used by so many vendors - was difficult for Koha. Need to grapple with that.
• Lookup User - need a good way to view whether a user / patron is active. 
• Need a way to charge / track payments in self-service machines (SIP2). SIP2-24 and SIP2-25 JIRAs on fees are already listed.
• Many of us need integrations more or less done by January 2020 or soon thereafter. 6-month lead time needed, at least.
• Alabama will start working on test integrations by Fall 2019.
• Should we set up separate integration test environments? Tricky due to all the variable factors. Too much thrash of the data structure.
  • Is there a way we can reduce the thrash on the data model with new versions, to minimize integration breakage? Q4 January 2020 - need to set that as a somewhat frozen data structure so other things can get done.

(Reporting)

How to build up the dataflow from Okapi to the LDP

The Reporting / LDP situation is challenged at the moment, not enough resources, much more work needed in designing the schema (a moving target with new releases of FOLIO)

Lead time: depends on how many integrations one institution needs to integrate

Jason: We need the things working in 6 month lead time / we need test instances and training etc. If that doesn't exist in the Jan 2020 release, we have to change to whole timeline.

Classic integration problems: have test instances ...

Lead time for Lehigh + Chicago : 6 month, too.

Tod: Defining export files. Unit testing would catch changes in the underlying modules. How do we allow for that ? But that does not minimize the sheer amount of work for integrations.