Slide 1
Slide 1 text
Writing Python Code
to Decide an Election
Peter Lubell-Doughtie
Slide 2
Slide 2 text
http://www.ly.undp.org/etc/designs/
UNDPIndiaDesign/img/home/logo-
undp.png
Slide 3
Slide 3 text
No content
Slide 4
Slide 4 text
No content
Slide 5
Slide 5 text
No content
Slide 6
Slide 6 text
the problem
• tally many votes from many districts in many
elections
!
Slide 7
Slide 7 text
the problem
• tally many votes from many districts in many
elections
• elections may span multiple district
Slide 8
Slide 8 text
the problem
• tally many votes from many districts in many
elections
• elections may span multiple district
• each ballot has a 7-stage verification process
Slide 9
Slide 9 text
Intake
Slide 10
Slide 10 text
Intake
Data Entry
Slide 11
Slide 11 text
Intake
Data Entry
Corrections
Slide 12
Slide 12 text
Intake
Data Entry
Corrections
Quality Control
Slide 13
Slide 13 text
Intake
Data Entry
Corrections
Quality Control
Archiving
Slide 14
Slide 14 text
…the other problem
• we had 3 weeks
Slide 15
Slide 15 text
…the other problem
• we had 3 weeks
Slide 16
Slide 16 text
only the essentials
Slide 17
Slide 17 text
only the essentials
• what functionality must exist?
!
Slide 18
Slide 18 text
only the essentials
• what functionality must exist?
• what tools already do some of that?
Slide 19
Slide 19 text
only the essentials
• what functionality must exist?
• what tools already do some of that?
• what if we miss something?
Slide 20
Slide 20 text
a schema as a tool
Slide 21
Slide 21 text
a schema as a tool
• decide how data is organized
!
!
Slide 22
Slide 22 text
a schema as a tool
• decide how data is organized
• shared reference
!
Slide 23
Slide 23 text
a schema as a tool
• decide how data is organized
• shared reference
• scopes possible relationships
Slide 24
Slide 24 text
a schema as a tool
• decide how data is organized
• shared reference
• scopes possible relationships
• it’s updatable
Slide 25
Slide 25 text
issues and milestones
Slide 26
Slide 26 text
issues and milestones
• convert discussion to issues
!
Slide 27
Slide 27 text
issues and milestones
• convert discussion to issues
• decide prioritization through milestones
Slide 28
Slide 28 text
issues and milestones
• convert discussion to issues
• decide prioritization through milestones
• communicate to what will get done when
Slide 29
Slide 29 text
do it
Slide 30
Slide 30 text
do it
1. assign an issue
!
!
!
!
!
Slide 31
Slide 31 text
do it
1. assign an issue
2. write tests for the issue
!
!
!
!
Slide 32
Slide 32 text
do it
1. assign an issue
2. write tests for the issue
3. write code for the tests
!
!
!
Slide 33
Slide 33 text
do it
1. assign an issue
2. write tests for the issue
3. write code for the tests
4. issue a pull request
!
!
Slide 34
Slide 34 text
do it
1. assign an issue
2. write tests for the issue
3. write code for the tests
4. issue a pull request
5. review the code
!
Slide 35
Slide 35 text
do it
1. assign an issue
2. write tests for the issue
3. write code for the tests
4. issue a pull request
5. review the code
6. pass the CI
Slide 36
Slide 36 text
do it
1. assign an issue
2. write tests for the issue
3. write code for the tests
4. open a pull request
5. review the code
6. pass the CI
7. merge and go to step (1.)
Slide 37
Slide 37 text
automate speed
Slide 38
Slide 38 text
automate speed
• anything that is done twice gets a script
Slide 39
Slide 39 text
automate speed
• anything that is done twice gets a script
• anything that breaks once gets a test
Slide 40
Slide 40 text
we used south
Slide 41
Slide 41 text
we used postgres
Slide 42
Slide 42 text
we used postgres
Slide 43
Slide 43 text
nose, flake8, and coverage
Slide 44
Slide 44 text
travis for ci
Slide 45
Slide 45 text
do not write code
Slide 46
Slide 46 text
do not write code
• unless you really really have to
!
!
!
Slide 47
Slide 47 text
guardian for permissions
Slide 48
Slide 48 text
enumfield state machines
Slide 49
Slide 49 text
enumfield state machines
Slide 50
Slide 50 text
reversion for data purity
Slide 51
Slide 51 text
fabric for remote execution
Slide 52
Slide 52 text
plan to handle what you
have not planned for
Slide 53
Slide 53 text
plan to handle what you
have not planned for
• elections can span multiple districts
Slide 54
Slide 54 text
plan to handle what you
have not planned for
• elections can span multiple districts
• did not know until the first day of live testing
Slide 55
Slide 55 text
plan to handle what you
have not planned for
Slide 56
Slide 56 text
plan to handle what you
have not planned for
• we changed the data model => south
!
!
Slide 57
Slide 57 text
plan to handle what you
have not planned for
• we changed the data model => south
• we wrote tests => django-nose
!
Slide 58
Slide 58 text
plan to handle what you
have not planned for
• we changed the data model => south
• we wrote tests => django-nose
• we changed the code => travis-ci
Slide 59
Slide 59 text
plan to handle what you
have not planned for
• we changed the data model => south
• we wrote tests => django-nose
• we changed the code => travis-ci
• we deployed => fabric
Slide 60
Slide 60 text
sage advice
• “simplicity is prerequisite for reliability” - Dijkstra
Slide 61
Slide 61 text
references
• http://south.readthedocs.org/
• https://github.com/5monkeys/django-enumfield
• http://www.fabfile.org/
• https://flake8.readthedocs.org/
• https://bitbucket.org/ned/coveragepy
Slide 62
Slide 62 text
thank you!
• https://github.com/onaio/tally-ho
• http://company.ona.io/
• http://peet.ldee.org/