We have been using OpenELIS lab system as part of the Bahmni Distribution. There are many feature requests from our customers but we couldn’t move fast due to its old software stack and ability to modify it confidently due to lack of code quality and good test coverage. This document tries to address some of these concerns and an opinionated view of next steps w.r.t making bahmni lab more flexible and customizable.
I would request you to please go through it and provide your comments/suggestions and if this approach really works. Thank you.
Interesting. I can see how this would definitely make it much easier to make improvements to our ELIS codebase.
It also guarantees that our forking from OpenELIS would be permanent. (Since we’ve been forked for 5 years, and never merged, maybe this is just reflecting reality.)
Looking at https://github.com/openelisglobal/openelisglobal-core it does seem like there’s some activity now. (I assume that’s the descendent of the thing we forked from svn…) It’s probably worth having a quick conversation with the current OpenELIS devs about this.
+1 on Mavenising its build (with Java 8), that’s a short term must have.
The users of OpenELIS are currently using 10% of the total software. (44 out of a total of 316 JSP pages)
That’s a very interesting input!
Under the assumption that the fork might become permanent, this really invites to look into a broader refactoring into a new and more lightweight app built on React (as you mentioned React in your doc.) I suspect that you have shied away from this idea for obvious scope and resources-related reasons.
My understanding in recent discussions with other devs of OpenELIS is that there is some new development happening on the software. There is interest in moving away from Struts, and decoupling the analyzer interfacing. I would encourage finding a way to collaborate with other OpenELIS devs before permanently forking, especially since Bahmni implementations are growing throughout Africa where OpenELIS is used in several countries. Maybe the community of practice is one method for this - since some of what might be desired could be leveraged by multiple open source LIS systems.