Stream: fmg
Topic: QA period
Grahame Grieve (Aug 02 2018 at 02:40):
We did not talk about QA this morning.. and there are still 100s of tasks to apply
Lloyd McKenzie (Aug 02 2018 at 03:02):
We don't have any formal QA baked into this ballot cycle. There's only one week between final freeze and ballot open. The logic at the time was that we probably wouldn't be reballoting too much and we'd have a significant QA cycle pre-publication. Plus that asking for too much QA in a year would be problematic.
Lloyd McKenzie (Aug 02 2018 at 03:03):
If there are specific high-priority things to QA between the 17 and the 24, we could try to get people to do that.
Grahame Grieve (Aug 02 2018 at 03:03):
what's here: https://chat.fhir.org/#narrow/stream/35-fmg-(members.20only)/subject/QA
Grahame Grieve (Aug 02 2018 at 03:04):
and I have raised some specific issues on the fhir-i stream
Lloyd McKenzie (Aug 02 2018 at 17:38):
There are a few additional STU voted items that appear to be on the wrong ballot. I'm going to interpret the FMG votes yesterday as allowing me to disposition those as Not Related rather than Not Persuasive on the reconciliation spreadsheets. Holler before Saturday if you have an issue with that...
Grahame Grieve (Aug 07 2018 at 12:47):
GF#15158 - this is a real squeaky wheel for me.... but before I take to this page... who owns it? FHIR-I, FMG, product Director?
Lloyd McKenzie (Aug 07 2018 at 14:15):
FHIR-I usually deals with the category stuff if we vote on it, but you tend to design it yourself regardless :) I'm not thrilled with the notion of putting all 'Definition' artifacts in the same place as the stuff they're defining.
Grahame Grieve (Aug 07 2018 at 20:39):
I wouldn't necessarily make a global rule either. But Q and Qr not being together remains odd for me and others
Lloyd McKenzie (Aug 07 2018 at 20:42):
The benefit of them being in separate spots is that you can find them regardless of which aspect you're looking at. I'd expect ObservationDefinition vs. Observation to work the same way.
Grahame Grieve (Aug 13 2018 at 13:19):
ok. we have a week for QA. The things I care about are:
- all resources listed on resourcelist.html (at least one is not)
- all pages correctly classified as to normative/trial use/etc
- all substantive changes listed for normative content
- trial notes all removed from normative content
- no inconsistent language left over describing past things (hanging chads!)
- unresolved todo items in normative content
- inconsistencies in definitions
- look for extensions that have been replaced by core elements but are still defined
- check that pages labelled informative do not include 'SHALL' or 'SHOULD'
Grahame Grieve (Aug 13 2018 at 13:20):
can anyone in FMG volunteer to take any of these on?
Grahame Grieve (Aug 13 2018 at 13:20):
oh: look for extensions that have been replaced by core elements but are still defined - that's for ballot now...
Grahame Grieve (Aug 15 2018 at 11:08):
report for FMG on the status of the ballot: the main FHIR spec is basically ready to go. I know of 4 open issues:
- OO discussion Observation.context
- terminology discussing a spelling error on a normative code (I want to talk about this during the FMG meeting)
- I have to update the UML diagrams
- I'm still working on the json schema
... + we need to ask committees to approve the content
IGs:
- plenty of work is happening, but I have not, as yet, had a chance to survey preparation
- at least one IG that has heaps of work is listed as cancelled (breast-cancer) - why is that?
- we need to set some earlier deadlines - some editors have suddenly started thinking about how to author a guide this week....
Lloyd McKenzie (Aug 15 2018 at 14:25):
So perhaps IG publishing on CI build by 1 month out? We could indicate that all artifacts must be present, but it'd be hard for us to know/track
Grahame Grieve (Aug 15 2018 at 21:23):
I added FMG approval status to http://build.fhir.org/resourcelist.html (on the maturity tab). It'll appear in about 30min. The results make for depressing reading.
Grahame Grieve (Aug 15 2018 at 21:49):
ok list is up. enjoy.... :cry:
Grahame Grieve (Aug 15 2018 at 21:50):
@Lloyd McKenzie and @David Hay : I added a section in fhir.ini and then copied/pasted from the approved and pending lists on the wiki.
Lloyd McKenzie (Aug 15 2018 at 21:58):
Did you take into account those that were grandfathered from STU2? Also, we did a bunch of the financial ones as a block.
Grahame Grieve (Aug 15 2018 at 22:07):
I just followed your method ;-) I don't know of anything in writing about grandfathering. But feel free to update the fhir.ini list
Lloyd McKenzie (Aug 15 2018 at 22:18):
Working on it
Lloyd McKenzie (Aug 15 2018 at 22:18):
Dealing with all of the renames and approved splits of approved resources
Lloyd McKenzie (Aug 15 2018 at 22:40):
I'm grandfathering Binary, Bundle, Composition, ImmunizationRecommendation as all have existed since DSTU1 - long before we had proposals. We can decide if we want to bother creating proposals for them
Lloyd McKenzie (Aug 15 2018 at 22:43):
Only two long-term resources that weren't in STU1 and don't have proposals is Parameters. There's a couple that are above FMM0 that don't have proposals, but they're pretty new and presumably should have proposals
Lloyd McKenzie (Aug 15 2018 at 23:02):
We didn't list any of the search parameters as normative - are they normative if the resource is normative? (Some of them probably shouldn't be as it's doubtful they've been used at all...
David Hay (Aug 15 2018 at 23:26):
So you're doing my task for me? Thanks!
Lloyd McKenzie (Aug 15 2018 at 23:44):
So items that aren't grandfathered and are either pending or have no proposals: OperationDefinition, Parameters, MessageDefinition
Lloyd McKenzie (Aug 15 2018 at 23:44):
OperationDefinition and Parameters we probably ought to approve given that they're candidate normative... :)
Grahame Grieve (Aug 16 2018 at 00:10):
Parameters should be grandfathered?
Grahame Grieve (Aug 16 2018 at 00:11):
search parameters have the same status as the element they are based on unless specified otherwise
Lloyd McKenzie (Aug 16 2018 at 00:26):
Parameters didn't exist in DSTU1, nor did OperationDefinition
Lloyd McKenzie (Aug 16 2018 at 00:27):
There's a lot of search parameters I'm sure have never been used in real systems (and might never even have been invoked on the reference servers)
Grahame Grieve (Aug 16 2018 at 01:11):
possibly. but we have no way of doing anything about this procedurally at this point
Lloyd McKenzie (Aug 16 2018 at 01:13):
Not in terms of what appears in the spec that goes to ballot, no. So I'd suggest we suppress that view for now - now that we've highlighted the issue - to avoid spurious ballots. We can formally approve the two important ones next week. And we technically don't have to have our ducks in a row until we publish in December.
Lloyd McKenzie (Aug 16 2018 at 01:14):
BTW - the approval view now looks correct based on all of the renames, splits, merges, etc. that have happened over time.
Grahame Grieve (Aug 16 2018 at 01:29):
suppress what view?
Lloyd McKenzie (Aug 16 2018 at 01:42):
"Approved by FMG"
Grahame Grieve (Aug 16 2018 at 02:04):
now I'm lost... what's the problem with the FMG approved list?
Lloyd McKenzie (Aug 16 2018 at 02:21):
It says that two normative artifacts aren't approved. That'll attract negative ballots, even though they will be approved before the deadline. And frankly, no-one will care about that view.
Lloyd McKenzie (Aug 16 2018 at 02:21):
What matters is the FMM 0
Lloyd McKenzie (Aug 16 2018 at 02:22):
Once we get things aligned properly, we can raise an error if FMM level is asserted to be anything other than 0 on a non-approved artifact.
Grahame Grieve (Aug 16 2018 at 02:33):
ok you want to remove the whole thing? the page is informative, so ballots are out of scope....
Brian Postlethwaite (Aug 16 2018 at 02:36):
Is the approved section of that table then not really needed? (as those are all the others)
And I agree as an informative page, its outside the ballot scope too.
Brian Postlethwaite (Aug 16 2018 at 02:37):
(Sorry for my absence this morning, not a great time to have a nasty cold)
Lloyd McKenzie (Aug 16 2018 at 02:40):
It'll draw attention anyhow. I see it as just noise. Who's going to care other than FMG members? If it's above 0, then it's FMG-approved (or our processes are broken, which they won't be soon). If it's 0, then it's not ready. Whether FMG-approved or WG hasn't fully committed to it, possibility of it going "poof" are about the same.
Grahame Grieve (Aug 16 2018 at 02:41):
maybe I'll move it to fmg.html and not link to it from anywhere
Lloyd McKenzie (Aug 16 2018 at 02:43):
That works. And add a todo to make it an error to be above FMM0 if not approved.
Last updated: Apr 12 2022 at 19:14 UTC
