2008-08-07 Developers Conference Call

Date

7 August 2008

In Attendance

  • Ben Wolfe
  • Brian McKown
  • Mark Benscoter
  • Burke Mamlin
  • Paul Biondich
  • Mike Seaton
  • Mohammad Shahiduzzaman
  • Darius Jazayeri

Agenda

  • Dave suggested adding an "empiric" column to the drug_order table
  • Intern review - Shahid
  • Unit Testing a-thon discussions

Minutes

  • Shahid is working on the Patient Note Writer with Andreas as his mentor
    • Chose tinyMCE as the wisywig editor
    • Working on two separate steps. image handler and editor
    • Notes are stored as observations
  • Putting off emperic discussion until Dave is on the call
  • Unit testing
    • Output of test a thon?
      • Documention about how to do a unit test?
      • Gold standard unit test section/
    • Proposal:
      • One developer creates first pass at gold standard with one of the small services
      • Set up a call with the one developer and the rest of the group. (everyone must be present)
      • At the testathon: pair off to work on tests?
      • At the testathon: solve the problem of lack of tests over the last 3 years vs the problem of creating an test driven developing environment
      • Pretestathon: figure out "priority": the "important" workflows and what "tasks" are done in that. Figure out what methods are involved in those and the specific behaviors within that and what should happen within the methods
      • We don't want to just test one per method...we want /at least/ one per method. Test each of the behaviors within the method with a different test method.
      • How do we know we're done?
  • Deliverables from testathon:
    • Documentation about how to do tests (Conventions as well)
    • Making writing tests easier to do later
    • Consensus about how to do testing
    • Test all of those prioritized methods/behaviors and while doing that find all of the edge use-cases
    • A unit testing roadmap
    • (a distant last) Test all api methods
  • Deliverables for pretestathon discussion (next week's conference call):
    • Straw man for one section (Alert, Encounter, Location)
    • Keep notes while writing the section as to what the "ah ha!"s were. this can be written up later
      • Write from the perspective of backfilling tests to the api
    • Look into mock objects?