Page MenuHomePhabricator

Step 1: count how many tainted reference warnings are being triggered and opened
Closed, ResolvedPublic8 Estimated Story Points

Description

Problem:
We want to know the following things:

  • How often a tainted reference icon is triggered (an edit is made where the main value changes and no reference changes)
  • How often someone clicks on the icon to see the popper
  • If our definition of tainted is a good one or should we also take into account:
    • If more all references should be changed
    • If qualifiers changing matters

BDD
GIVEN a UI edit
WHEN the value is changed
AND no snak of any of the reference is changed
THEN it is counted as one statistic

GIVEN UI edit of a statement
WHEN the value is changed
AND at least one snak of every reference is changed
THEN it is counted as one statistic

GIVEN UI edit of a statement
WHEN the value is changed
AND at least one snak of a reference (but not all) is changed
THEN it is counted as one statistic

GIVEN UI edit of a statement
WHEN the value is not changed
AND at least one snak of a reference is changed
THEN it is counted as one statistic

GIVEN UI edit of a statement
WHEN the value is not changed
AND at least once qualifier snak is added, removed, or changed
THEN it is counted as one statistic

GIVEN a tainted reference warning
WHEN the icon is clicked to open the modal
THEN it is counted as one statistic

Acceptance criteria:

  • dashboard exists that tracks the number of triggered warnings and openings over time

note: all of the above should only need to happen when the feature flag is turned on.

Event Timeline

Tarrow subscribed.

Moving back to needs work since that's what we seemed to say in the last story time.

An additional statistic I mentioned was counting when some but not all references for a given statement remained unchanged on value change. I think (but I'm not sure) if we also thought there were other stats we wanted to grab

Lydia_Pintscher renamed this task from count how many tainted reference warnings are being triggered and opened to Step 1: count how many tainted reference warnings are being triggered and opened.Oct 11 2019, 10:30 AM
Tarrow updated the task description. (Show Details)

Here are some examples for how we interpret this ticket:

All references changed statistics should be fired in the the following examples:

Case 1
BEFORE
Main Statement: Population -> 1000
References:
  - Book A on page 3
  - Book B on page 4


AFTER
Main Statement: Population -> 2000
References:
  - Book C on page 5
  - Book D on page 6
Case 2
BEFORE
Main Statement: Population -> 1000
References:
  - Book A on page 3
  - Book B on page 4


AFTER
Main Statement: Population -> 2000
References:
  - Book E on page 7
Case 3
BEFORE
Main Statement: Population -> 1000
References:
  - Book A on page 3
  - Book B on page 4


AFTER
Main Statement: Population -> 2000
References:
  - Book E on page 7
  - Book F
  - Book H

Some references changed:

Case 4
BEFORE
Main Statement: Population -> 1000
References:
  - Book A on page 3
  - Book B on page 4


AFTER
Main Statement: Population -> 2000
References:
  - Book A on page 3
  - Book E on page 7
Case 5
BEFORE
Main Statement: Population -> 1000
References:
  - Book A on page 3
  - Book B on page 4


AFTER
Main Statement: Population -> 2000
References:
  - Book A on page 3
  - Book E on page 7
  - Book F on page 8
Case 6
BEFORE
Main Statement: Population -> 1000
References:
  - Book A on page 3
  - Book B on page 4

AFTER
Main Statement: Population -> 2000
References:
  - Book A on page 3
  - Book B on page 4
  - Book C on page 5

Note Case 6 is a little unusual. This might need an extra category in the data collection. Perhaps we should have one specific statistic for "no references removed but some added". What do you think @Lydia_Pintscher ? Right now we lump them all in with the "some references changed" stats

As per daily: @Lydia_Pintscher says leaving as is is fine :)