2026-08-13 Developers Deep Dive Call - Roles
Overview
Date: 2026-08-13
Time: 10:00 am (UTC)
Room: https://meet.jit.si/openIMISDevelopers - kindly indicate your name and organisation (jitsi claims to support Chrome, Chromium & MS Edge and Firefox browsers only. You might also want to try the jitsi app)
Participants: (kindly only add your own names, not those of other participants)
@Uwe Wahser
m4h , Y-Note , Tinker Technologies
@Sunil Parajuli
Agenda:
When | Duration | Who | Topic |
|---|---|---|---|
10:00 am |
|
| roles in openIMIS |
Issues with roles
https://drive.google.com/drive/folders/1-hxdxvDDLQ2y-4LiWtgKx4GB7ng8bmXR
Wiki documentation: Access management
Options for improvement
customizable rules (Cameroon)
introduce roles not with only access right (right->roleRights in openIMIS context) but sometime we want is_medical_officer, or is_claim_admin , for example if we want to hide the ‘health facility page’ to Medical Officer, we need to override or work on rights (example: 100011), that will disable the page view for ‘Admin’, ‘Medical Officer’ (as they shared same rights). So, if we have to disable to role associated with “is_medical_officer” then we can disable that menu specifically for this kind of user keeping it more simple.
fixes need to be done one-by-one - not common root cause visible
we need an approach for efficient detail testing of all the roles
impersonification would help / maybe only on test-servers
how do we adress all the flagged roles?
is there a root cause or do we need to adress one by one?
depends on where the issue is: front-end vs. back-end
LLM support test against backend in CI → task sent to SolDevelo
Frontend will be difficult , as (if backend sends correct ids for rights but user are not able to do the operation even they are are allowed) → analysis in Bugfixing contract
way forward:
LLM can be done quite quickly → wait for the results before adressing frontend
fix backend, at the same time develop testing strategy for frontend
Analyse frontend related issues after cleaning backend
Questions
Attachments
Additional Resources
PrismML Releases Bonsai 27B: 1-bit and Ternary Builds of Qwen3.6-27B That Run on Laptops and Phones - MarkTechPost https://share.google/2OFWa7OohPVRkpnRo
https://github.com/hackathon-develop/2026-07_healthcare-hack_openimis.git