Stream: JIRA/Confluence
Topic: Voting vs. transitioning issues
Lloyd McKenzie (Jan 17 2020 at 22:24):
There are a number of places where we have votes and dates filled in, but the issues still has a status of "Triaged". Once you have a vote you should always be transitioning the issue to Resolved - Change Required, Resolved - No Change, etc. To help with this, I've removed the ability to specify a vote and date when using the "propose a resolution" screen - so you'll always have to open the desired transition screen in order to specify a vote. However, if doing a bulk change, you'll still need to remember to perform the transition once the votes have been filled in.
Also, a reminder that you may need to 'move' the issue to the appropriate category before resolving. If something is a question, you can't find it persuasive or non-persuasive without first moving it to be a change request. Similarly if you determine that a technical correction is substantive, it must also be moved to be a change request.
If there are any questions/issues, please let me know.
Michael Donnelly (Feb 04 2020 at 03:52):
Who should fix these? Does it have to be a co-chair, or is it appropriate for anyone to do it when it's clear what the status should be?
Michael Donnelly (Feb 04 2020 at 03:53):
(I'm specifically looking at https://jira.hl7.org/browse/FHIR-21230)
Lloyd McKenzie (Feb 04 2020 at 04:00):
Anyone who's been granted permissions as a "trusted user" can do so
Michael Donnelly (Feb 06 2020 at 05:24):
I did a search for trackers that have a resolution but are still in a status of Triaged. Saved search here: https://jira.hl7.org/issues/?filter=12836
Michael Donnelly (Feb 06 2020 at 05:29):
There were more than 1300 of those. I was able to clean up a couple hundred of them, but there are a bunch that have one or more of the following problems:
- The resolution isn't valid for the issue type
- e.g. you can't find a question persuasive
- Confluence has the valid type/resolution pairs
- No vote was recorded (in the Resolution Vote field, at least)
- It has a Change Impact and/or Change Category when it shouldn't
- It doesn't have a Change Impact and/or Change Category when it should
Michael Donnelly (Feb 06 2020 at 05:58):
This is how many each workgroup has:
| Workgroup | Invalid Resolution for Type | Valid Resolution, Missing Resolution Vote | Valid Resolution, Wrong Impact and/or Category | Total |
|---|---|---|---|---|
| Biomedical Research & Regulation | 2 | 19 | 3 | 21 |
| CIMI | 0 | 13 | 12 | 13 |
| Clinical Decision Support | 12 | 132 | 29 | 169 |
| Clinical Genomics | 0 | 9 | 3 | 9 |
| Clinical Interoperability Council | 6 | 75 | 22 | 81 |
| Clinical Quality Information | 0 | 3 | 3 | 3 |
| Community-Based Care and Privacy | 0 | 2 | 1 | 2 |
| Devices | 0 | 1 | 1 | 1 |
| FHIR Director | 0 | 1 | 1 | 1 |
| FHIR Infrastructure | 4 | 51 | 45 | 55 |
| FHIR Mgmt Group | 1 | 0 | 0 | 1 |
| Financial Mgmt | 9 | 233 | 152 | 323 |
| Modeling & Methodology | 0 | 2 | 2 | 2 |
| Orders & Observations | 2 | 30 | 14 | 32 |
| Patient Administration | 1 | 4 | 2 | 5 |
| Patient Care | 1 | 52 | 24 | 53 |
| Pharmacy | 0 | 2 | 2 | 2 |
| Public Health | 61 | 151 | 169 | 266 |
| Publishing | 1 | 0 | 0 | 1 |
| Security | 1 | 3 | 3 | 4 |
| Structured Documents | 0 | 7 | 6 | 7 |
| US Realm Task Force | 3 | 0 | 0 | 3 |
| Vocabulary | 0 | 27 | 10 | 27 |
Michael Donnelly (Feb 06 2020 at 06:00):
Here's a spreadsheet with all the issues: Jira-Trackers-Resolved-but-still-Triaged.xlsx
Michael Donnelly (Feb 06 2020 at 06:02):
Though I'd recommend just doing your own search in Jira. You can start with Voted on but not resolved FHIR trackers and then add on additional search terms.
Michael Donnelly (Feb 06 2020 at 06:04):
For example, I see that BR&R has 19 issues where there's a resolution but no vote. To find those, I'd start with the search linked above and append AND "Work Group" = "Biomedical Research & Regulation [brr]" AND "Resolution Vote" is EMPTY.
Michael Donnelly (Feb 06 2020 at 06:06):
That returns 21 results because it also brings in two questions (which can't be Persuasive or Persuasive with Mod). So I add on
AND type = "Change Request"
and get the 19 trackers that are missing their votes.
Michael Donnelly (Feb 06 2020 at 06:09):
@Lloyd McKenzie has updated Jira to prevent issues from getting into many (or maybe all?) of these states in the future, so this won't be an ongoing cleanup task.
Lloyd McKenzie (Feb 06 2020 at 09:57):
@Michael Donnelly Those with no vote recorded aren't necessarily an issue so long as they are marked as "ready for vote" or one of the "block vote" groups and even without that, they might be preliminary dispositions awaiting conference call discussion. The other two columns are only problematic if they have votes. (If they don't have votes, the work groups will have to fix the issues before they can add the vote information and transition them.) How much work would it be for you to re-do this analysis with those points in mind?
Michael Donnelly (Feb 06 2020 at 23:07):
How much work would it be for you to re-do this analysis with those points in mind?
To a fair extent, it's already there, and adding other stuff won't be too hard. The sheet includes a column for Resolution Vote, so people can filter on that. That's definitely worth doing first:
- It's a much smaller set of issues.
- Like you say, they're definitely issues when many of the others aren't.
Michael Donnelly (Feb 06 2020 at 23:07):
I'm going to crunch a bit and report back.
Michael Donnelly (Feb 06 2020 at 23:23):
Here are the ones that have a vote recorded:
| Workgroup | Invalid Resolution for Type | Valid Resolution, Wrong Impact and/or Category | Total |
|---|---|---|---|
| Clinical Decision Support | 12 | 25 | 37 |
| FHIR Mgmt Group | 1 | 0 | 1 |
| Financial Mgmt | 6 | 81 | 87 |
| Orders & Observations | 1 | 0 | 1 |
| Patient Administration | 1 | 0 | 1 |
| Public Health | 51 | 54 | 105 |
| Publishing | 1 | 0 | 1 |
| Grand Total | 73 | 160 | 233 |
Pivot tables!
Michael Donnelly (Feb 06 2020 at 23:24):
Those are ones that someone who knows more than I do might be able to resolve without input from the WGs. I looked at a bunch of them, but I wasn't confident of what should be done.
Michael Donnelly (Feb 06 2020 at 23:32):
Unfortunately, there definitely are problems with some of the issues that don't have votes. For example, look at the history of J#17593. It was voted on, and had its status set to Considered for Future Use. Then a few months later on a date PA didn't have a meeting, someone changed the resolution to Persuasive, presumably by mistake. This wiped out some of the other fields.
Michael Donnelly (Feb 06 2020 at 23:51):
I don't know how much we can do to detect that programatically. Maybe something that hits the API and looks for weird stuff in the history?
Michael Donnelly (Feb 06 2020 at 23:52):
Likely it would be less total work for people to review them.
Michael Donnelly (Feb 06 2020 at 23:53):
@Lloyd McKenzie If an issue is prepped for a block vote, or if someone has set a preliminary disposition, should the resolved timestamp be set?
Jean Duteau (Feb 07 2020 at 01:56):
@Michael Donnelly Thanks for this thread. I've gone and created some JIRA filters for my two WGs so that I could find these "orphan" issues where we've proposed a disposition but forgot to vote on that disposition. Been very helpful
Lloyd McKenzie (Feb 07 2020 at 02:57):
Resolved timestamp gets set as soon as there's a candidate resolution. It's not a terribly meaningful element for us - not one we have control over
Michael Donnelly (Feb 07 2020 at 03:54):
Hmm. What element(s) can we look at to determine whether an issue with a resolution but no Resolution Vote is in a problematic state?
Lloyd McKenzie (Feb 07 2020 at 05:44):
I don't think something with a resolution without a vote is a problematic state. Work groups should be worrying about anything that's Triaged or Waiting for Input - whether it has a resolution or not.
Lloyd McKenzie (Feb 07 2020 at 05:45):
If it hits anything past 'Triaged' without a vote and it's not a Technical Correction, that's an error.
Brian Postlethwaite (Feb 07 2020 at 22:56):
We often add the disposition ahead of time while preparing for calls.
Was that PA issue you highlighted the one from the table? (as it's now fixed) - so I don't have to go Hu ting if we have another out there. Thanks again for the analysis too.
Last updated: Apr 12 2022 at 19:14 UTC
