Wikimedia Engineering/2014-15 Goals/Q1

Wikimedia Engineering Top Priorities, FY2014-15, Q1 Q2

Top departmental priorities for Q1 (July - September 2014) edit

Top 5 user-impacting priorities edit

Objective Impact Dependencies ETA Status (end of quarter)

1) Migrate to HHVM as execution environment for PHP code

Performance for editors. Expected to halve median backend response time for uncached requests, including preview and save operations, increasing user engagement and productivity.

Metrics

Dept-internal

  • MediaWiki Core
  • Operations
  • Developers across engineering (socialization)

End of Q1

Mostly completed. Production rollout delayed by 4-6 weeks (proceeding in stages in October-November)

2) Launch new native apps for Android and iOS. Iterate on contributory features, including new mobile-only features on mobile web

Expected to build a stable cohort of mobile app focused contributors, but exact impact unknown - will establish baseline for app contributions/reach vs view.

Metrics

Dept-internal

  • Apps team/Product/UX
  • Partnerships (preloads, featured app)

Dept-external

  • Communications (rollout)

Launch beginning of Q1, then iterating on functionality through Q1

Fully completed. Native apps launched, first WikiGrok experiments in testing.

3) Deploy Flow to significant Wikipedia use case (in consideration: Teahouse or similar workflow)

Expected to validate Flow as an effective discussion system (compare basic participation metrics before/after; UX testing); we can begin establishing a activity/retention baseline to improve from.

Dept-internal

  • Flow team/Product/UX/Community

Dept-external

  • Potentially help from Teahouse team (community and staff experts)

End of Q1

Partially completed. Flow used (as advertised experimental alternative) for French Wikipedia new user help forum. First usage metrics reported in quarterly review.

4) Continue improving VisualEditor (citations, performance, visual parameter-editing, IE support); begin re-engaging with English Wikipedia community

  1. New citations functionality will make citations - currently obscure, template-based machinery - truly accessible to anyone by simply pasting a URL. This will benefit wikitext users as well.
  2. VisualEditor is expected to increase -- measure before/after activation/retention/bounce rate of new users. These are critical path improvements to make this an actually realistic goal -- as is making VisualEditor available to all users on English Wikipedia.

Metrics

Dept-internal

  • VE/Product/UX/Community

Dept-external

  • Recruiting (2 front-end devs)

End of Q1

Partially completed + work on unscheduled functionality. IE11 support released; client-side compression released; Citoid puppetized and in Labs; work has begun on reviewing unanticipated downstream contribution of a table editor.

5) Consolidate remaining detached user accounts into Single User Login system ("SUL finalization")

Enables cross-wiki functionality to be built without exception handling. Significant immediate impact on power users whose accounts are affected, needs to be carefully managed. Complete all technical work to zero-account conflict and a plan for roll-out.

Dept-internal

  • Platform/Community/Product

Dept-external

  • None

End of Q1

Mostly completed (some last technical work outstanding, see SUL finalization; expected to be completed at a lower priority in October).

Top 5 platform/process improvement priorities edit

Objective Impact Dependencies ETA Notes

1) Migrate to HHVM as execution environment for PHP code

See above

See above

See above

See above

2) Conduct pre-launch pilots with Phabricator and launch "day 1" installation including Bugzilla import

Increase developer efficiency and mitigate silo effects by reducing tooling fragmentation (code review, bug tracking, project management)

Dept-internal

  • Operations/Platform collaboration, socialized across all of engineering

Dept-external

  • Guinea pigs welcome!

End of Q1

Mostly completed, preparing for Bugzilla migration in October.

3) Improve MediaWiki Core’s front-end libraries

Reduce wheel reinvention and other silo effects across teams; begin shift towards multi-device development; build proper UX standardization development pipeline

Dept-internal

  • Primarily VE/Mobile engineers, affecting short term velocity

Ongoing, but first deliverables to be scoped out for Q1

Q2 work fully completed, ongoing.

4) Standardize baseline metrics & dashboards

Enable more consistent measurement of user-impacting changes for Editors, Readers and Content. [1]

Dept-internal

  • Analytics, Product

Dept-external

  • Collaborate w/ Grantmaking

End of Q1 for Editors; TBA for Readers, Content

Fully completed, see quarterly review.

5) Launch the new Content API, prepare for use by Mobile in Q2

Medium term: Bring logged-in user pageview performance on par with logged-out user performance. Long term: Significantly improve load/save performance in VisualEditor by minimizing round-trips.

Dept-internal

  • Services, Parsoid, Operations, Platform, Mobile collaboration

Launch in Q1, Mobile use in Q2

Partially completed, running in Labs.