2014-04-03 Developers Forum
How to Join
Agenda
Quickly review previous meeting minutes (5 min)
Release Process
OpenMRS 1.11
After-action review & next week's agenda (5 min)
In Attendance
You
Minutes
OpenMRS Developers Forum: 2013-04-03
Recording: https://connect.iu.edu/p51udotdun3/
Attendees
Michael Downey
Roger Friedman
Karl Wurst
Nyoman Ribeka
Wyclif Luyima
Burke Mamlin
Elliott Williams
Daniel Kayiwa
Ada Yeung
Darius Jazayeri
Rafal Korytkowski
Mhawila Mhawila
Joseph Kaweesi
Agenda & Minutes
Review last week's TODO's - http://notes.openmrs.org/Developers-Forum-2014-03-27
TODO: Burke will get PIH presentation from Mike Seaton added to the wiki page
TODO: Burke & Darius to clean Road Map up
TODO: Start having more hackathons - Discuss at https://talk.openmrs.org/t/lets-plan-a-mid-year-developer-event/ (ongoing)
TODO: Send out an e-mail announcement to dev@ & implementers@ informing people about the new project. (Elliott)
Release Process
Reviewing the current release process
Timing of releases (i.e., setting some goals)
We never want to go more than a year between releases
Releasing an update (minor or major) to our reference application at least twice a year
Marchs & Septembers (e.g., merge code in Jan & July)
Releasing an update (minor or major) to our platform application at least once a year
Septembers (e.g., merge code in July)
When do we set "the date" for a release? Feature-driven or time-driven... or compromise?
We'll try time-based in the 2014-2015
Recipe for releasing (being a release manager)
Why isn't anybody taking notes here?
Need to consider separate release process page for reference application vs. platform?
Need more automated tests for OpenMRS 2.0
TODO: Testing discussion in dev forum
Need to get basic release testing automated & sustainable
Still just taking notes by myself. /me wishes we were better at taking notes during meetings.
s/etherpad/wiki/g (note that Burke has been better about copying & pasting notes into wiki after meetings)
TODO: We need to find way(s) to encourage implementations to help test new releases
What things could make it worthwhile? Are there any incentives?
Release Testing Module was a good idea, but was stopped b/c of issues with getting module-specific data transferred.
Could modules support a testing interface (getTestData & saveTestData)?
Are there out-of-the-box approaches we could take to encourage testing?
Can we create demo sites with different combinations of modules?
We currently have 1 system for UAT on order and more can be obtained.
Plan on ordering 72 of them.
Dream on!
Ok. We can settle for 30.
Mike has built some pretty cool tools for exporting random/helpful subsets of data, deidentifying them, etc. There was also a SoC project about this sort of thing. => If OpenMRS can set up a hosted server with all the relevant modules, and we just need to get one implementation to send us realistic testing data.
TODO: really could use a realistic demo data set – i.e., great to have the tools, but where are the de-identified data?
Implementations probably feel more comfortable giving deidentified test data with limited access, just to the implementation itself plus 1-2 trusted OpenMRS devs/testers.
Contrived test data is safe for crowdsourced testing