Technology

Annual Plan Collaboration Jam Meeting

22-24 February 2017

Technology

Cross-Departmental Segments:

• Privacy and Security
• New Readers
• Structured Data
• Verifiability

Other programs Technology is expected to provide consulting to include:

• Anti-harassment
• Movement strategy

Departmental Programs:

• Develop Technology’s audiences
• Cloud Services
• Measure the adoption of Wikimedia Technology
• Tech debt repayment
• Streamlined delivery of services
• Scoring platform (ORES)
• Reduce content gaps via recommendations
• Diversify the contributor population
• Renewed focus on the MediaWiki platform

Cross-Departmental Program Segments

Security, Privacy and Data Management

​

​

​

We seek to maintain a strong security and privacy posture for the Foundation and the movement. To accomplish this goal, several departments within Technology, along with Legal, will work together to ensure that privacy standards, security process, and data collection and storage techniques meet the needs of the organization and community.��Segments Team Leads: Security, Analytics

• Outcome 1: Improve organizational security posture
• Objective 1.1: Increase capacity to participate in security-centric activities
• Objective 1.2: Address vulnerabilities resulting from external security assessment
• Objective 1.3: Update tools and processes to keep pace with industry-wide security developments�
• Outcome 2: Data is accessible in a way that is consistent with the privacy expected by our users and the values of the movement
• Objective 2.1: Guide process for creation/description of new datasets
• Objective 2.2: Ensure retention guidelines are being followed
• Objective 2.3: Better offboarding / onboarding for data access
• Objective 2.4: Anonymization for newer public safe datasets, e.g. editors per country or pageviews per country

​

​

New Readers

New Readers is bringing new people into the Movement, people who don’t yet have a strong understanding of our projects or our mission. By creating better experiences and access for those potential readers, we will enable them to access free knowledge content. Education is a right. Design Research collaborates with Global Reach, Reading, Communications, Community Resources, toward these ends. More readers around the world will also mean more contributors, especially among people, languages, and cultures that are underrepresented in the wikis now. ��Segments Team Leads: Design Research, Services

• Outcome 1: Solution sets across all 3 segments are grounded in user needs and constraints. Ideas come from generative research and are validated by evaluative research.
• Objective 1.1: Evaluative studies (usability testing + concept evaluation) of software solutions
• Objective 1.2: Workshops for concept generation and evaluation across teams
• Objective 1.3: Generative research as needed�
• Outcome 2: The New Reader initiative's API and infrastructure needs are well designed and implemented.
• Objective 2.1: Work with Reading to address API and infrastructure needs for the New Reader initiative deployment.

​

Structured Data

We will introduce the concept of “multi-content revisions” to MediaWiki, allowing structured data to be associated with revisions of wiki pages, alongside traditional free-form text. This will allow structured data to be associated with images and other media on Wikimedia Commons, in collaboration with WMDE.

Segment Lead Team: MediaWiki Team

• Outcome 1: Enable extensions to efficiently store multiple content blobs associated with each revision.
• Objective 1.1: Modify the MediaWiki database schema
• Objective 1.2: Update the page save API
• Objective 1.3: Update the page retrieval and rendering API�
• Outcome 2: Enable extensions to provide a user interface for editing and reviewing multi-content revisions
• Objective 2.1: Adapt the edit and diff pages
• Objective 2.2: Update MediaWiki’s review and abuse prevention tools

Verifiability

The Wikimedia movement builds its reputation on our volunteers’ ability to vet and compile information from reliable sources. To support this goal we will design a number of initiatives to: (1) define metrics and methods for assessing how verifiable content on a subset of Wikimedia projects is and (2) increase editor and reader engagement around sources and citations.

�Segment Lead Team: Research

• Outcome 1: Readers and contributors can use new tools to vet and source information
• Objective 1.1: Establish new collaborations to conduct research and development on algorithmic strategies for verifying and assessing the reliability of Wikipedia articles and Wikidata statements.
• Objective 1.2: Conduct user research to understand how readers and contributors engage with sources and references. Help Product teams design interfaces to facilitate sourcing and verification of content.
• Objective 1.3: Collect, analyze and sanitize clickthrough data for footnotes and external links (after discussing and reviewing privacy and security implications)�
• Outcome 2: Community efforts on verifiability are aligned and coordinated
• Objective 2.1: Host the 3rd annual meeting in the WikiCite series. Promote it with research, open science and open access partners. Previous events in 2016 and 2017, funded via restricted grants, have produced significant impact.

Departmental Program Highlights

Develop Technology’s audiences

Since the 2015 reorganization of the Engineering department, the Foundation has been focused mainly on three audiences: readers, editors and new users (discovery). This has left a few less major, more tech savvy but nevertheless important audiences underserved: tool authors, API and dataset downstream consumers and third-party users of our software. All of these have been supported by the Technology department's teams on a best effort basis and with a limited understanding of their needs, their size and diversity and with no way to measure our progress in serving them. In FY 2017-2018 we will implement a set of initiatives to expand and strengthen these audiences, improve our engagement with them, as well as consistently measure the outcome of those efforts. Team Participants: Cloud Services, Research and Data, Design Research, ORES, MediaWiki, Analytics, Services.

• Outcome 1: Strengthen our Cloud Services products and improve productivity of their users
• Objective 1.1: Reduce confusion about the differences between Cloud Services products
• Objective 1.2: Provide improved technical support and documentation�
• Outcome 2: Ensure the adoption of Wikimedia technology can be reliably measured
• Objective 2.1: Design a set of formal KPIs to measure the growth and diversity of our technology audience.
• Objective 2.2: Host an annual thematic event targeted at our audience of technology, API and data consumers �
• Outcome 3: Strengthen our support to third-party users
• Objective 3.1: Establish canonical point of contact for third-parties
• Objective 3.2: Clarify Foundation’s short- and long-term commitments to third-party users

Cloud Services

Wikimedia Cloud Services (WMCS) will support the technical contributors to the Wikimedia movement by providing a stable and effective hosting platform for technical projects relevant to the Wikimedia movement when such projects are incompatible with Wikimedia Foundation production hosting. Additionally we provide technical support, documentation, and community support to foster growth and productivity of the community of users of the Cloud Services products. Team Participants: Cloud Services, Technical Collaboration, Communications

​

• Outcome 1: Reduce confusion about the differences between the 'Labs' and 'Tool Labs' products
• Objective 1.1: Complete fiscal year outlined rebranding activities by 2017-12-31�
• Outcome 2: Users are able to find documentation they need. Agree/Disagree answer ratio for "Documentation is easy to find" survey question raised to 65/35.
• Objective 2.1: Provide first line technical support resources to triage and respond to Cloud Services product support requests.
• Objective 2.2: Form documentation special interest group to update documentation of existing Cloud Services products.
• Objective 2.3: Tutorial content created for common issues of creating initial account(s), deploying a functional web service, deploying a functional bot, running a periodic job with variations for more than one implementation language.�
• Outcome 3: Implement modern redundancy and stability features in OpenStack which can be exposed for customer benefit
• Objective 3.1: Transition networking topology to OpenStack Neutron

Measure the adoption of Wikimedia technology

The Technology group’s audience consists of individuals, internal teams, movement affiliates and organizations that rely on services, APIs, datasets and technology we build. The success of our mission depends on the continuous growth and engagement of this audience. Right now, we have a limited understanding of this audience, how large and geographically diverse it is, and how much progress we’re making in serving it. In FY 2017/18 the Technology group will implement a set of initiatives to consistently measure the growth of this audience and to improve how we engage with it. Team Participants: Analytics, Cloud Services, ORES, Design Research, Research and Data, Services

• Outcome 1: Adoption of technology produced by the Wikimedia Foundation can be reliably measured
• Objective 1.1: Design a set of formal KPIs to measure the growth and diversity of our technology audience. Reference these KPIs in our quarterly and annual progress reports.�
• Outcome 2: Technology, API and data consumers have a dedicated forum
• Objective 2.1: Host an annual thematic event targeted at stakeholders of the technology group. Promote their engagement in identifying technical directions and priorities and showcase applications they are building based on our technology offer..

Tech Debt Repayment

To increase the overall velocity of the entire Wikimedia developer community, we will create, and follow through on, a tech-debt repayment plan. We need a way to address the endemic structural issues with our code bases that slow down our ability to increase stability and release new features. To maximize initial effort of this program we will focus on 1) code being used by Wikimedia to serve our users and 2) WMF teams and projects driven by WMF teams. Team Participants: Release Engineering, Team Practices Group

• Outcome 1: Reduce the amount of orphaned code that is running Wikimedia “production” services.
• Objective 1.1: Define a set of code stewardship levels (from high to low expectations)
• Objective 1.2: Identify, and find stewards for, high-priority/high use code segment orphans
• Objective 1.3: Define and steward a light-weight process for adopting or orphaning/sunsetting products and infrastructure.�
• Outcome 2: Reduce organizational technical debt
• Objective 2.1: Define a “Tech Debt PM” role that regularly communicates with all WMF engineering teams regarding their technical debt
• Objective 2.2: Define and implement a process to regularly address technical debt across the WMF
• Objective 2.3: Promote and surface important technical debt topics at large gatherings of Wikimedia developers (eg DevSummit and Hackathon(s))

Streamlined delivery of services

​

We are building a new production platform for integrated development, testing, deployment and hosting of applications. This will greatly reduce the complexity and speed of delivering a service and maintaining it throughout its lifecycle, with fewer dependencies between teams and greater automation and integration. The platform will offer more flexibility through support for automatic high-availability and scaling, abstraction from hardware, and a streamlined path from development through testing to deployment. Services will be isolated from each other for increased reliability and security.

Wikimedia developers as well as third party users benefit from the ability to easily replicate the stack for development or their own use cases. Team Participants: Technical Operations, Release Engineering, and Services�

• Outcome 1: Seamless productionisation and operation of (micro)services
• Objective 1.1: Set up production-ready Kubernetes cluster(s) with adequate capacity
• Objective 1.2: Create a standardized application environment for running applications in Kubernetes�
• Outcome 2: Developers are able to develop and test their applications through a unified pipeline towards production deployment.
• Objective 2.1: Create guidelines and abstractions for building and testing applications in containers
• Objective 2.2: Set up a CI pipeline to publish new versions of an application to production via testing and staging environments that reliably reproduce production
• Objective 2.3: Provide a lightweight integrated development environment that lets developers test their code against a local miniature copy of the production stack

Scoring Platform (ORES)

AI has great potential to help our projects scale by reducing the work that our editors need to do and enhancing the value of our content to readers, but AIs also have the potential to perpetuate biases and silence voices in novel and insidious ways. ORES, is a high capacity, machine learning prediction service that is already heavily adopted within and outside the Wikimedia Foundation. In the next fiscal year, we’ll create a dedicated team to further develop ORES and related technologies to balance efficiency and accuracy with transparency, ethics, and fairness. � Team Participants: Scoring Platform, Research, Technical Operations, Services

• Outcome 1: Tool devs and Product can innovate tools that use machine prediction to wiki-work easier
• Objective 1.1: Expand vandalism & good-faith detection models to more wikis (focus on Emerging Communities)
• Objective 1.2: Develop new types of useful predictions (e.g. “draft quality” and “edit types”)�
• Outcome 2: Volunteers are empowered to track trends in prediction bias and other failures of AI in the wiki
• Objective 2.1: Build technologies for enabling community-based auditing of predictions (“effective refutation”)

Reduce content gaps via recommendations

We aim to further develop algorithms and methodologies that can help Wikimedia editors identify what to edit next with the goal of closing Wikimedia knowledge gaps.

​

Team Participants: Research, Technical Operations, Services

• Outcome 1: Increased (in variety) and improved recommendation services for editing Wikimedia projects
• Objective 1.1: Develop and test methodology(/ies) for article (stub) expansion via recommendations
• Objective 1.2: Build APIs to surface recommendations

​

• Outcome 2: Provide recommendations on other Wikimedia projects
• Objective 2.1: Build a recommendation API for another Wikimedia project, such as Wikidata, Wiktionary, or Wikimedia Commons

Diversify the contributor population

The goal of this research program is to diversify the contributor population to Wikimedia projects. Diversifying includes, but is not limited, to increasing gender diversity in the editor population.

​

Team Participants: Research, Resources

​

• Outcome 1: Test the impact of building more confidence on increasing the diversity of editor population
• Objective 1.1: Recommendations (supported by experiments) that can improve the diversity of editor population on Wikipedia (See Voice and Exit in a voluntary work environment)
• Outcome 2: Perform funnel analysis to assess the diversity of contributors population at each stage of the contribution funnel

​

Renewed focus on MediaWiki platform

Instructions: Add a short description of your program, including the Goal, here. This is intended as an exercise in conciseness. Below, list the Outcomes, and Objectives that your segment or team(s) are planning. Keep it brief; use multiple slides if necessary. Team Participants (Departmental Program):

• Outcome 1
• Objective or Activity 1.1
• Objective or Activity 1.2
• Etc.
• Outcome 2
• Objective or Activity 2.1
• Objective or Activity 2.2
• Etc.

Resourcing

Resourcing & dependencies

Resourcing & dependencies