2011-02-10 Developer Meeting
Date
10/Feb/2011
How to Join Call
In Attendance
@Ben Wolfe
@Andrew Kanter
@Wyclif Luyima
@Paul Biondich
@Stephen Lorenz
@Jeremy Keiper
@Burke Mamlin
@Dawn Seymour
@Darius Jazayeri
@Diptanu Choudhury
@Carl Fourie
@Glen McCallum
@Roger Friedman
@Mark Goodrich
@Daniel Kayiwa
@Nyoman Ribeka
Agenda
Quickly review previous meeting minutes (5 min)
OpenMRS 1.8 Update (5 min)
OpenMRS 1.9 Planning (10 min)
Stanford all night Hack-a-thon - Stanford Hack-a-thon 2011 (5 mins)
Concept Mapping Improvements -- TRUNK-412 (45 mins)
Clarification of release process & release manager responsibilities - Burke et al. (45 mins)
Review of De-duplication (patientmatching) module (15 mins)
After-action review & next week's agenda (5 min)
Minutes
1.8 update:
Wyclif:
Set to have the beta 2 to release friday. Will not include the 'intitialization wizard database copier ticket'.
Darius: working on multi-column index ticket. Will not be in 1.8 but instead in a module that can be done at any point.
Roger: please explain the module decision
Darius: Having it in a module allows impl to run it at any time.
Roger: I proposed having it as a liquibase in main line with precondition that if they have less than X observations.
Roger and Darius will take discussion offline
Burke: Do we need an RC?
Ben: Yes, but only if there aren't any changes in the beta
1.9 planning:
Reviewing tickets on roadmap.openmrs.org:
Multiple tickets are partly done.
concept mapping will be finally discussed today
Burke dislikes the 'documentation of interaction points with ancillary systems' ticket.
Paul: This ticket is largely for documentation. Some code exists, some doesn't. This is mostly for things like 'I have a lab integration system, how would I integrate?'. Would be nice ot have a web page for that
Darius: this might involve some code being written.
Diptanu: I can help contribute to that document. We use openmrs and other systems
Burke: I want at least something on a wiki page instead of a bunch of points on a call that are lost.
Shaun: is this a list of 'you should do this this and this' or is it a best practices?
Burke: Its more of a pointer to people that have done things
Burke: there is not '1 hammer' to solve everyone's needs
Stanford Hack-a-thon
A handful of student hackers will be working on openmrs tickets for 24 hours straight this weekend.
We need people to help them on irc/forum, especially overnight on Saturday Feb 12th
Concept Mapping Updates
concept mapping improvements for 1.9
Andy: We also need a derived table for exploding children/grandchildren/etc for speed improvements
Burke: We can do that later potentially
Andy: But if it is not usable with the speed improvements, its useless
Roger: concept map type might not need to be a user defined table
Burke: openmrs does not want to be an ontology management system. modules can add their own types.
Darius: using a table adds a lot of complexity
Burke: fair enough, but we shouldn't assume that the values are in X enum and will never have more
Roger: what about the open archetypes standard
Carl: We're working on that and the occ integration to share concepts
Roger: Andy has mapped his concepts in a snomed way. Is this data strucutre sufficient to handle OSA model. If not, is it possible to make somethign to handle both
Burke: it'd be nice to let other systems do what they do best, openmrs can just provide the basics
Roger: can some of this be in core vs a module?
Andy: The relationship type of tables can probably be a module
Andy: There could be entries in the refernece_term table that are not mapped to a concept yet.
Roger: translations are not possible in the term table
Andy: we weren't looking to create a dynamic translator here
Andy: do we want a source column in the term_map table?
Glen: what if the terms in the term_map are for two different source?
Darius: just join to term to get the source
Andy: how to know which rows in termp_map to delete when doing an update?
Glen: do a pattern matching on ... ____
Darius: terms have a unique id. Do term_maps have a code?
Andy/Glen: Yes, they probably do
Andy: could add a 'external_code' column to term_map' (TODO)
Darius: How to work if we add locale to reference_term (TODO)
Darius: How does this work if term_map has a code? (TODO)
Ben: can we rename concept_map to something more in line with concept references
Burke: concept_map ? concept_reference_map
Jer/Roger: need some documentation on how this works with concrete examples
Roger: Darius' weight example is a good one to use