Phlogiston/2016-04-25 collaboration

From mediawiki.org

Joe, Joel, Roan.

Reviewed http://phlogiston.wmflabs.org/col_report.html to see what we can learn prior to the Collab team retro tomorrow. Three things stood out as meaningful from the data

Several categories not likely to finish this quarter as currently scoped[edit]

http://phlogiston.wmflabs.org/col_backlog_burnup_count_zoom.png shows that much less than 1/3rd of the backlog is done, although we are 1/3rd into the quarter. And, the backlog continues to grow with no leveling suggested.

Individual burnup charts show the same thing. Forecasts suggest rate of completion is behind rate of growth.

Options for cutting and prioritizing:

  • Don't accept new tasks. If you accept a new task, remove an old task.
  • Identify a bundle of work that is both high-priority and only valuable if all tasks are completed. Do all of that before doing anything else
  • Identify some sub-categories that can be prioritized with a category.

Individual tasks do seem to be small enough[edit]

Based on http://phlogiston-dev.wmflabs.org/col_done_count.png (works on phlogiston-dev, may be broken on phlogiston), Collab is closing 10 to 15 tasks per week. With ~5 dev equivalents this would be 2-3 tasks per week, which seems to be a reasonable balance.

"Other" is clogging up some charts[edit]

http://phlogiston-dev.wmflabs.org/col_recently_closed.html

Phlogiston can't differentiate between tasks that actually finished last quarter or earlier and were just now resolved as Phab cleanup, and actual work completed in the last week. This chart could indicate whether or not the team is working on low-priority stuff, but it can't do that if clogged up by Phab cleanup. No obvious, cheap fix other than to not clean up old work, so doing anything different because of this may not be worth it. but be aware of the value of prompt Phab cleanup. Zombie Phab tasks have other costs beyond affecting Phlogiston reports. Could check this again in a few weeks to see if it's cleared up.

Bugs/Issues[edit]

- nominal and optimistic velocity should be higher for Q4 Flow forecast

- Q4 Notifications Work burnup is too pessimistic