FHIR Chat · Call minutes · ig publishing requirements

Stream: ig publishing requirements

Topic: Call minutes

Lloyd McKenzie (Jul 06 2021 at 23:40):

Minutes of our initial call are here: https://confluence.hl7.org/display/FHIRI/FHIR+IG+Authoring+Minutes+CC+20210706
Thanks to all who participated.

Lloyd McKenzie (Jul 13 2021 at 22:21):

Minutes from today's call are here: https://confluence.hl7.org/display/FHIRI/FHIR+IG+Authoring+Minutes+CC+20210713

Lloyd McKenzie (Jul 20 2021 at 21:11):

Only Jose & I there today, so we called it after 10 minutes. (We contemplated just writing up and voting to approve the process while consensus was easier, but figured that wouldn't get us very far :smile: )

Bob Milius (Jul 20 2021 at 21:29):

I showed up about 12 minute late. Sorry about that. :sad:

Lloyd McKenzie (Jul 21 2021 at 01:49):

np. We'll try again next week.

Eric Haas (Jul 21 2021 at 22:16):

I was on too late as well sorry

Lloyd McKenzie (Aug 03 2021 at 22:05):

Minutes from today's call are here: https://confluence.hl7.org/display/FHIRI/FHIR+IG+Authoring+Minutes+CC+20210803

Lloyd McKenzie (Aug 10 2021 at 23:47):

Minutes from today's call are here: https://confluence.hl7.org/display/FHIRI/FHIR+IG+Authoring+Minutes+CC+20210810

Jose Costa Teixeira (Aug 11 2021 at 06:03):

Sorry I missed it.
On the templates / current, I guess my idea was a separation between current (latest released) and CI build releases - like we have in the core spec and in IGs. I don't know if that is the target or whether it should be.

Lloyd McKenzie (Aug 11 2021 at 13:35):

If you reference the template without a version, you'll get the most recent release.

Lloyd McKenzie (Aug 24 2021 at 22:00):

Minutes from today's call: https://confluence.hl7.org/display/FHIRI/FHIR+IG+Authoring+Minutes+CC+20210824

Grahame Grieve (Aug 24 2021 at 22:19):

I have some comments on this:

Grahame Grieve (Aug 24 2021 at 22:19):

this line:

impact the IG-Publisher in a way that changes what IG authors can (or may) do or what IG readers see

Say someone points out that the IG publisher is producing output that is wrong. Fixing that is a fix to something that IG readers see, but should not require this process to be followed with all the sign off. It should be for a change to what IG readers are intended to see

Second issue is similar: we regularly get IG authors (Lloyd!) asking for changes to allow something new far down in the edge of what we're doing, and it won't impact anyone else. the IG publisher team should be allowed to create new edge case functionality that is reasonably expected that won't impact on anyone else without going through the envisioned review process

Same for fixing bugs in e.g. snapshot generation, that impact on IG authors

Also, the definition is obtuse with regard to some of the most significant changes we make for HL7 IG authors: increasing validation that causes new entries to appear in qa.html

Lloyd McKenzie (Aug 24 2021 at 22:38):

Agree that something that meets the definition of a 'technical correction' (obvious error with equally obvious fix) shouldn't require approval process.

For stuff on the edge, I think it depends on how likely it is that others will also hit that edge eventually. For stuff like 2016May-specific changes, where there's only one implementer in the world, agree seeking feedback doesn't make sense. If introducing something not yet used but that eventually many might use, seeking discussion on how it should work at the outset is probably still appropriate.

Fixing the validator to detect things that are true errors or warnings according to the spec wouldn't need this process. Adding on additional layers of QA beyond what the specification mandates would presumably come through this process.