Milestones

Platform & Backend Improvement Ideas for 2024

• Audit Log for Views & Updates: We have no auditing around views. Updates: if a form is filled out multiple times, we'd only have information on the first time and the most recent update. Large project. (eg auditlog work stagnant) There are workarounds, eg Statistics Module, or by configuring the server to log URLs requests (Tomcat or the entry proxy). Wyclif Luyima recommends that we use https://hibernate.org/orm/envers/ instead of https://github.com/openmrs/openmrs-module-auditlog
• Investigate Batching Suggestion from Implementers: Optimize backend requests via batching?
  • (Grace & Daniel follow-up on Palladium's issue here) - eg Explore Support for GraphQL https://graphql.org/ to organize request batching
  • Where do we need to batch? Would batching actually solve the issue?
• Optimize slow REST API calls performanceperformance
• Age Range support
• Med List
• Patient Consent support

Platform 2.7 (Q1 2024 - before mid 2024)

Have feedback about this list? We want to hear from you! Please share in the forum thread here: https://talk.openmrs.org/t/openmrs-platform-roadmap/38652 

Our main goal: Make Platform as easy to support as possible.

1. DONE on core, IN PROGRESS with modules - Support for Java 17 - many newcomers come with Java 17 installed, run into errors, we discover the issue is Java 17. RefApp currently runs well on Java 11, but not 17. (LTS after 11 is 17.) People should be able to develop in the latest LTS of Java, but we want to be able to run on a previous LTS, e.g. still support running on Java 8 (thinking of Implementers). (As we try out 17, will depend on what we figure out; if can't support Java 8, might decide to either cancel 17 update, or drop 8, etc. - we try not to just drop support for a previous version due to the difficulty this causes for implementers.)  TRUNK-6197. As part of this work, the reference application modules need to be updated to also support Java 17, but without dropping support for Java 8.
2. IN PROGRESS: Upgrade to Liquibase 4.19.0 - for patches included. Need to check if this upgrade would cause any complexity for us. TRUNK-6155, TRUNK-6208
3. Upgrade other small libraries to their latest versions. Dependabot seems to be doing a good job for us here.
4. DONE: A REST endpoint to check status of platform (especially when platform is starting up) TRUNK-6198. Would be great to have sense of progress for long-running states.
5. DONE for core, TODO for Modules: Remove all uses of the Hibernate Criteria API because it’s deprecated (replaced with JPA criteria) TRUNK-6202
6. TODO: Global properties access should be privileged, so we need an endpoint to expose those items needed (anonymously) during O3 startup. TRUNK-6203, TRUNK-6206
7. DONE: Add support for init parameters for module servlets (TRUNK-4673) and this Talk post
8. TODO: Docker as a first-class approach to deployment. RA-1990, TRUNK-6186, TRUNK-6083
9. ON HOLD - Implement Encounter Context and Visit Context. Much of encounter context work may apply to frontend; there may be some requirements for business on the backend (e.g., deciding when to use existing vs. create new encounter).
   1. (Pending FE guidance): an approach that lets applications discover/add to "open encounters" for a patient rather than collecting notes, orders, observations, etc. and then trying to infer which or how to group them properly into clinical encounters. This is a need for OPD and adding enough Java dev support to that effort to propose (e.g., to community & TAC) how to better organize data into clinical encounters.. A /dev/4 or /dev/5 with Java (backend) experience & familiarity w/ the model. Probably needs input from clinical SMEs too.
      • E.g. if I directly add an allergy or add a new condition, which encounter is it supposed to be logged in? 
      • E.g. I need to enter lab results that came in after the patient left the clinic - or results from weeks ago that the patient has just handed me, while  they also have an active visit.
10. ON HOLD - Extract Immunization like we did for Allergies, Conditions, Diagnosis (Was not done). Need to be able to attribute these (including Allergies, Conditions, Diagnoses) to an encounter. (On hold - lack of consensus)

Platform 2.6 (Q1 2023)

Platform 2.5 (Q4 2021)

• Support for Tomcat 8.5+
• Support for FormRecordable for more than just Obs and Condition 
• Administration via REST

Tickets for Platform 2.5

Additional candidates for 2.5:

• Patient Lists (Cohort Module changes)
• User settings

  Jira Legacy
  server OpenMRS Issues columns key,summary,type,created,updated,due,assignee,reporter,priority,status,resolution serverId 45c5771b-fa4b-3e43-b34a-c19dc45ccc95 key TRUNK-6020

• Moving OrderTest toward ServiceOrder to support Referrals

2021 Goal Brainstorming

Actual Tech Priorities (Ideas):

• Platform QA: Some regressions - need more automation to catch in advance, esp. build into PR pipeline
  • Automation +++: Get platform to point where enough is automated so have confidence in quality and release more frequently
• Simplifying Deployment
• Support for timezones across the Platform (and beyond, see also about HFE here.)
• FHIR
  • Support for FHIR Group through maybe a new Core Group entity replacing Cohort (see Slack discussion here.)
• Metadata Sharing: a solution that meets majority of needs (if Iniz, maybe expand and handle a few additional use cases; come up with something that's better than our old tech for metadata sharing)
• Integration: Clarify actionable things we can do to better integrate into other Digital Health tools that are typically found around an EMR like OpenMRS.
• i18n: Fully internationalizing metadata concepts (but is this really a priority, pain point?)
  • Support for i18n of metadata in Core (see for example this thread.)
• Download & Run: Easier to use without getting bogged down in details - something non-tech end user can work with. Standalone hasn't advanced in a while - needs love for feature updates or different approach or make useable again (e.g. some Mac versions no longer running). People using Standalone in production b/c simple to manage, came up with backups automation approach. → So easy to download, click and it runs. Using tech that's already available on users' computers. 
• Sync: Most implementations not connected to single server; most combine, do analysis later. Is this a priority for implementers now?
• Patient Lists: FHIR-based patient lists proposal from Ampath
• Library dependencies between modules (not high priority because so time consuming, and there are (ugly) workarounds)
• Decision Support: Some meaningful step (e.g. with simpler ETL solution: flattened data, generates table, SQL, extrapolate from that). Feels we are waiting for AES to eventually address.

Release Calendar

How to update this calendar

List View

Calendar View

Someday

Reference Application Someday

Platform Someday