openmrs plans/ideas 3.0.0

Development
1

Hi there!

I just wanted to share some things I think might improve OpenMRS which might be bigger changes so I thought of the roadmap and saw 3.0.0. Not sure if there is already a place where there are wishes/plans. Wiki is currently down so I cant see https://wiki.openmrs.org/display/docs/Technical+Road+Map

Some thought are:

  • make openmrs cloud-native/adhere to 12 factor apps rules https://12factor.net/ (like for example take configs out of environment variables not files/java properties; only log to console appender not files). This would make it easy to deploy openmrs on cloud providers, ease development
  • ensure spring starts immediately when openmrs is deployed (and remove all the static Context, OpenmrsUtil methods). Leverage Spring to get configs out of the environment. Make use of spring profiles?
  • use java 8 time api, instead of java.util.Date

Are these realistic for example for 3.0.0? What would need to be done for the openmrs ecosystem (modules, distributions, implementations, …) I am sure you have tons of ideas of what you would you like to see, please share! And what can we do to already work towards these goals

2

Did you mean to say, take configs out of environment variables or the opposite? :slight_smile:

3

Thanks @teleivo for bringing this up!

Since we have stumbled on the issue of making OpenMRS/Bahmni support a full language switch (see Ext I18N), I would add:

  • Enable I18N support for metadata names and descriptions (on the model of what’s done with the Concept class).
4

thats what I mean (not the opposite :slight_smile:)

5

@teleivo but the link you gave seems to say the opposite. :slight_smile:

6

Not sure where it says that. On The Twelve-Factor App it states:

The twelve-factor app stores config in environment variables

7

Spring boot might also be interesting, creating a runnable openmrs with embedded app server:

8

I was looking at your “take configs out of environment variables” vs “The twelve-factor app stores config in environment variables”.

9

Well it goes both ways :laughing: the app stores and takes/gets its config in/from environment variables.

10

I think I see the confusion:

“take config out of environment variables”

  • What @teleivo meant: “obtain configuration through environment variables”

  • What @dkayiwa read: “remove configuration from environment variables”

Thanks to English for making it too easy to take ambiguity out of language. :slight_smile:

11

Love these ideas. Some more thoughts for OpenMRS 3.0.0 (some technical, some more cultural):

  • Add more business logic into the Platform (e.g., incorporating emrapi methods like we discussed recently in a design forum).
  • Steps toward cluster support
    • Remove use of singletons (as you mentioned)
    • Begin using & promoting redis (or similar) for application state
    • Review queries for those that might not be safe in a clustered enviromment & refactor them
  • More automation, including integration tests
  • Utility functions for escaping all user-entered content by default when rendering (i.e., opt-out of escaping when displaying content) to better address XSS vulnerabilities.

Stretch goals:

  • Convert to postgres as default database
  • Migrate from JIRA to GitHub for issue management

…

For 4.0.0:

  • Make our API horizontally scalable (clusterable)

For 5.0.0:

  • Separate the API into microservices
  • Containerize services so they can leverage whatever technology suits their needs best (e.g., nosql databases)
12

Loool, @burke you got it man!!! :smile: