Wikimedia Engineering/2014-15 Goals/Q2

Q1 Wikimedia Engineering Top Priorities, FY2014-15, Q2 Q3
Objective Beneficiary[1]→Measure of success for Q2[2] Project Lead[3] Dependencies ETA Status

Understand new mobile web micro-contributions and mobile app readership

Conservative measures of success:

Readers→ [Wikipedia App] Improved searching/quick content browsing.

  • Additional search features (in-page search, Wikidata descriptors in search)
  • Improved top-of-page experience (lead image presentation, infoboxes, disambiguation & cleanup templates, etc.)
  • Consistency of article lead appearance

Success measured by usability testing and other qualitative user research.

New/Casual Editors→ [Mobile Web] Understand what factors produce better engagement and quality for WikiGrok, the first mobile micro-contribution feature, before testing in production.

  • Test (quantitatively and qualitatively) WikiGrok in beta on the following dimensions: target audience (logged in vs. logged out), number of articles, types of statements (at least 3 different statement types), UI (two different interaction types), point of entry (in-article or sidebar) – to see how these impact, usability, engagement and quality
  • Establish preliminary baseline for number of submissions per article to aim for in testing in stable next quarter and decide which other factors we need to manipulate (audience, number of articles, types of statements) for the test
  • Experiment with aggregating multiple submissions before pushing to Wikidata

Stretch measures of success:

  • [Mobile Web] A/B test of WikiGrok in stable to more accurately assess which UI produces best engagement-to-quality ratio. We are blocked on unstable backend infrastructure (Tool Labs/Wikidata) for releasing WikiGrok more widely.
  • [Mobile Web] Launch prototype of second mobile micro-contribution feature in alpha/beta. We may be able to pilot a second Wikidata game project due to increased engineering resources on the mobile web team.
  • [Wikipedia App] Show measurable increase in reader engagement via at least one new reader-facing feature. We are blocked on apps readership data, lack of iOS developers, and not getting adequate design resources until one month into Q2.

Maryana Pinchuk

Product, UX, Engineering, Analytics

End of Q2.

Longer-term plan (Q3-4) here.

Yes Done

Stretch goals mostly missed, exception: WikiGrok was tested in stable with logged in users.

Make editing faster (for Q2: focus on instrumentation)

Conservative measures of success:

Developers→Instrumented edit funnels on desktop/mobile/app, wikitext/VE

Stretch measures of success:

Editors→# 50th/99th percentile load/save perf.

Editors→% activation/retention

Ori Livneh

Engineering, Analytics, Operations

Likely 'til end of Q3

Incomplete Partially done

  • Optimistic saving was implemented for the wikitext editor, reducing median save time (across the board) by 300ms.
  • Shared schema for logging performance metrics was created.
  • Instrumentation code for this was added to VisualEditor.
  • A custom backend for processing these metrics and routing them to Graphite was written and deployed.

Standardize and modernize front-end libraries

Conservative measures of success: (by end-Q2)

Developers→Elimination of dependencies on Mantle templates (MediaWiki core feature) and JS Object utilities (OOjs)

Developers→Integration of OOjs UI in JS/PHP into MediaWiki core, with at least one key interface converted as an example

Developers→Improved documentation of OOjs UI with examples, tutorials and reference information

Stretch measures of success: (by end-Q2)

Developers→Several key MediaWiki core interfaces converted as examples

Developers→Use OOUI components in the creation with Design of a Human Interface Guidelines document

Trevor Parscal

Engineering, UX, Product

Likely 'til end of Q3

Yes Done

Stretch goals missed, though the OOjs UI based living style guide is in active development.

Componentize MediaWiki core libraries

Developers→Move logging to composer managed library

Developers→Extract one library from existing MediaWiki codebase, publish as a stand alone library and re-import to MediaWiki via Composer

Developers→Complete documentation and implementation of Composer managed libraries for use on WMF cluster RFC

Developers→Create list of future candidates for library extraction projects

DevelopersInvestigate implementing a library management system for JavaScript libraries used by MediaWiki Deferred in consultation with frontend standards group

Bryan Davis

Engineering

End of Q2

Yes Done

Phabricator as Wikimedia software project management tool

DevelopersThe majority of WMF developer teams and sprints have moved to Phabricator

DevelopersCommon project management guidelines to be followed by teams and individual contributors are published

Quim Gil

Engineering, Operations

End of Q2

Yes Done

Executive Focus Areas

edit
  • Product & Strategy:
    • Document product development methodology and process, including criteria for deployment/rollout of functionality, policy for site-wide technical changes, and inclusion of community requirements and needs
    • Support first phase of strategy development process: community consultation, finalization of framework

Notes

edit
  1. Should refer to one of the Personæ or user categories.
  2. Measures of success to be fleshed out by project leads
  3. New role, discussion ongoing; see Wikimedia Engineering/Project Lead role