2020-11-20 - Sys Ops & Management SIG Agenda and Notes

Separation of "Reference Data" and "Custom Data" (Jo had another name for this) and "Sample Data" - Status of the work

Upgrades with Reference data

Wayne, Tod & Stephen have been in discussions about this for some time now in a working group.  

Settled on an adaptation modeled on the wiki page.  Will change terminology.

Starting to call it "default data", both required and optional.

Distinct from "operational data"

The basic architecture they propose is a module supplies default data available for a tenant that is using that module.  The data and module are both versioned.

Once loaded, it becomes operational.

The upgrade process will involve some form of reconciliation between the default data of previous module, new module and operation data created by the tenant over the course of operation.

Still a work in process.

Can we switch between languages in the reference data?  This does not cover internationalization.  Out of scope for the charge.  We don't have a good mechanism for this in FOLIO.

Place examples here:

Ann-Marie Breaux: "At today's sprint demo, David Crossley demonstrated a spot where he is gathering developer documentation automatically. Seems like it could be useful for the data migration and sys ops folks at FOLIO libraries. His presentation starts at ca. the 1:00 mark on the recording (which will be posted to Youtube shortly)."

Links:

• Source code map
• Presentation in FOLIO Sprint Review 100-1001

So far we said this:

• The group agreed that objectives must be measurable.
• The schedule for the Initiatives for mid 2021 is a too aggressive schedule. If not everything happens, we don't want it to feel like a failure.
• (enabling consortia support in a single tenant, and cross-tenant consortia support): SysOps think that one has to re-work the whole database to achieve this. So this cannot be a goal for mid 2021.
• (Initiatives & Objectives): We can't tell which initiative goes with what objective.

• We find the Initiatives too prescriptive. It suffices to have the Objectives and let the groups define the initiatives.

• (Clearly document the library requirements for stability (Mean Time Between Failures [MTBF], Mean Time To Repair [MTTR])  and secure realization) : It will be up to SysOps to define the requirements.
  
• (Objective for mid 2022 ) : We will have little control over this.
  
• (Strategic Objectives 3-5 years out): These should be our goals starting now. The project should be constantly working towards these. A 5 year plan is too far out. Our vision should carry us as we visit our objectives yearly and plan initiatives to support those objectives. We should work on this now. If we wait 3 to 5 years, FOLIO would have fallen behind the library industry and might not ever catch up!
• Don' wait with ensuring to have a capable funding team.

These are the various thoughts that were brought up during the discussion:

• The power point doesn't provide enough detail to provide a response based on understanding.
• We need more information.
• The power point is too high level; The levels of responsibility inside the group are too different to create a common response.
• The change is too slow.

We have dissenting opinions :

• There are people in the group that fully agree to the Governance Model.
• The change is too slow. We have to open the seats immediately.
• Re-work seats so that anyone is eligible. Do we even need a transition in the seat model ?
• Being too restrictive might be a fork in community.
• Who is the ultimate authority ? Not one council, but where are the checks and balances. In the current model, we don't see the checks and balances.
• The CC shouldn't have that much power; should be on equal parts.
• If there is a transition, that transition should happen much quicker.

In an open source project, the entity that provides the resources ultimately makes the decisions.

Does the rest of the world knows that Governance model ?

Current proposal could discourage institutions from participating

Koha is different. There are no ... councils. They don't have a board. It's open source.

Do we even need a transition in the seat model ?

The change is too slow. We have to open the seats immediately.