Wikilambda/Plan

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

This is a one page version of the individual sections which are transcluded. Click on the titles to go the individual pages.

Summary[edit]

This proposal is a draft and it is expected to see significant changes based on decisions and discussions with the communities and other stakeholders. Which, in turn, also means that comments and discussions are more than welcome in order to improve and shape the proposal.

In Abstract Wikipedia, Content is represented in a language-independent format which is directly editable by the community. The local Wikipedias can access and enrich their own locally controlled content with Content from Abstract Wikipedia. This way, the local Wikipedias can with much less effort provide much more content to their readers, content that is more comprehensive, more current, and more vetted than what most local Wikipedias can provide.

Given research results and prototypes in the area of natural language generation, it is unfortunately true that natural language generation from the language-independent Content requires a Turing-complete system. But in order to cover the number of languages Wikipedia needs to cover, this system must be crowdsourced. Therefore we introduce Wikilambda, a project to create, catalog and maintain a library of functions, which has many possible use cases. The main use case is the development of Renderers that turn the language-independent Content of Abstract Wikipedia into natural language, using the linguistic and ontological knowledge available in Wikidata.

The project will start with creating the Wikilambda project, and then use this in order to enable the creation of Abstract Wikipedia. Wikilambda will launch within the first year, and in the second year we add the development of Abstract Wikipedia. After two years, we will have created an ecosystem that allows for the creation and maintenance of language-independent Content and the integration of this content within the Wikipedias, significantly increasing coverage, currency, and accuracy of many individual Wikipedias. This will move us dramatically closer to a world where everyone can share in the sum of all knowledge.


Name[edit]

All names - Abstract Wikipedia and Wikilambda - are preliminary and meant mostly for writing this proposal and discussing it.

The current names are based on the following idea:

  • Abstract Wikipedia: the content in Abstract Wikipedia abstracts away from a concrete language
  • Wikilambda: this is based on the notion that all functions can be grounded in en:lambda calculus. Also, Wλ looks kinda geeky.

Note that the name "Abstract Wikipedia" will, in fact. not stick around. When the project is done, Abstract Wikipedia will be just a part of Wikidata. This is just a name for the development work, and therefore naming is not that crucial. Wikilambda on the other hand would be a new Wikimedia project, and thus the name will have rather high visibility. It would be good to come up with a good name for that.

There have been three good reasons against the name Wikilambda brought up so far:

  • It is really hard to spell for many people (says Effeietsanders)
  • Some people keep misreading it as Wikilambada (says Jean-Frédéric)
  • It also easily misreads as WikiIambda / Wikiiambda (that's, with yet another i / I instead of the l / L), so it should at least be WikiLambda with a capital L (suggested by Fuzheado)

Alternative names that have been considered or suggested:

  • WikiSetta or WikiRosetta (by BugWarp)
  • Multilingual Wikipedia
  • Translingual Wikipedia (by Deryck Chan)
  • Project Eco (to honor Umberto Eco’s work, in particular the book “The search for a perfect language”, which details the history of projects who have failed at the goals of this project)
  • Project Aquinas (which is more a play of word on the term "knowledge acquisition" than the philosopher Thomas of Aquinas)
  • Wikidata Abstract Content
  • Wikifunction for the Function repository
  • UniversalPedia (as per ChristianKl)
  • Wikiwords (as per Zblace)
  • Wikigram or Wikiprogram

Other suggestions are welcome.

In fact, the first task P1.1 for the project will be to decide with the community together on a name and on a logo. This had precedence for previous projects (Logo of Wikidata, Name of Wikivoyage).


Goals[edit]

The project has ambitious primary and a whole set of secondary goals. The primary goals are:

  • Allowing more people to read more content in their language
  • Allowing more people to contribute content for more readers, and thus increasing the reach of underrepresented contributors

Secondary goals include, but are not limited to:

  • Reusable and well-tested natural language generation
  • Allowing other Wikimedia communities and external parties to create content in more languages
  • Improving communication and knowledge accessibility well beyond the Wikipedia projects
  • Develop a novel, much more comprehensive approach to knowledge representation
  • Develop a novel approach to represent the result of natural language understanding
  • A library of functions
  • Allowing to develop and share functions in the user’s native languages, instead of requiring them to learn English first
  • Allowing everyone to share in functions and to run them
  • Introducing a new form of knowledge asset for a Wikimedia project to manage
  • Introducing novel components to Wikipedia and other Wikimedia projects that allow for interactive features
  • Create functions working on top of Wikidata’s knowledge base, thus adding inference to increase the coverage of Wikidata data considerably
  • Catalyzing research and development in democratizing coding interfaces
  • Enabling scientists and analysts to share and work on models collaboratively
  • Share specifications and tests for functions
  • The possibility to refer to semantics of functions through well-defined identifiers
  • Faster development of new programming languages due to accessing a wider standard library of functions in Wikilambda
  • Defining algorithm and approaches for standards and technique descriptions
  • Providing access to powerful functions to be integrated in novel machine learning systems

The list is not exhaustive.


Organization[edit]

We assume a core team, employed by a single hosting organization, that will work exclusively on Wikilambda and Abstract Wikipedia. It will be supported by other departments of the hosting organization, such as HR, legal, etc.

The team will explicitly be set up to be open and welcoming to external contributions to the code base. These may be volunteers or paid (e.g. through grants to movement bodies or by other organizations or companies). We aim to offer volunteers preferred treatment in order to increase the chances of creating the healthy volunteer communities that we need for a project of such ambition.

The project will be developed in the open. Communication channels of the team will be public as far as possible. Communication guidelines will be public. This will help with creating a development team that communicates publicly and that allows to integrate external contributions to the code base.


Requirements[edit]

The following strong requirements are based on the principles and practises of the Wikimedia movement:

  1. Abstract Wikipedia and Wikilambda will be Wikimedia projects, maintained and operated by the Wikimedia Foundation. This mandates that Abstract Wikipedia and Wikilambda will follow the principles and guidelines of the Wikimedia movement.
  2. The software to run Abstract Wikipedia and Wikilambda will be developed under an Open Source license, and will depend only on software that is Open Source.
  3. The setup for Abstract Wikipedia and Wikilambda should blend into the current Wikimedia infrastructure as easily as possible. This means that we should fit into the same deployment, maintenance, and operations infrastructure as far as possible.
  4. All content of Abstract Wikipedia and Wikilambda will be made available under free licenses.
  5. The success of Abstract Wikipedia and Wikilambda is measured by the creation of healthy communities and by how much knowledge is being made available in languages that previously did not have access to this knowledge.
  6. Abstract Wikipedia will follow the principles defined by many of the individual Wikipedias: in particular, Neutral Point of View, Verifiability, Notability, and No Original Research.
  7. Abstract Wikipedia and Wikilambda will be fully internationalized, and available and editable in all the languages of the Wikimedia projects. Whether it is fully localized depends on the communities.
  8. The primary goal is supporting local Wikipedias, Wikidata, and other Wikimedia projects, in this order. The secondary goal is to grow our own communities. Tertiary goals are to support the rest of the world.
  9. The local Wikipedia communities must be in control of how much Abstract Wikipedia affects them. If they don’t want to be affected by it, they can entirely ignore it and nothing changes for them.

The developers of Abstract Wikipedia do not decide on the content of Abstract Wikipedia, just like the developers of MediaWiki do not decide on the content of Wikipedia. Unlike with the other Wikimedia projects, the developers will take an active stand in setting up and kick-starting the initial set of types and functions, and creating the necessary functions in Wikilambda for Abstract Wikipedia, and helping with getting language Renderer communities started. Unlike with other projects, the development team of Abstract Wikipedia and Wikilambda will be originally more involved with the project, but aims to hand all of that over to the communities sooner rather than later.

The following requirements are used as strong guidances that we apply in the design and development of Abstract Wikipedia:

  1. Abstract Wikipedia and Wikilambda are a socio-technical system. Instead of trying to be overly intelligent, we rely on the Wikimedia communities.
  2. The first goal of Abstract Wikipedia and Wikilambda is to serve actual use cases in Wikipedia, not to enable some form of hypothetical perfection in knowledge representation or to represent all of human language.
  3. Abstract Wikipedia and Wikilambda have to balance ease of use and expressiveness. The user interface should not get complicated to merely cover a few exceptional edge cases.
  4. What is an exceptional case, and what is not, will be defined by how often they appear in Wikipedia. Instead of anecdotal evidence or hypothetical examples we will analyse Wikipedia and see how frequent specific cases are.
  5. Let's be pragmatic. Deployed is better than perfect.
  6. Abstract Wikipedia and Wikilambda will provide a lot of novel data that can support external research, development, and use cases. We want to ensure that it is easily usable.
  7. Wikilambda will provide an API interface to call any of the functions defined in it. But there will be a limit on the computational cost that it will offer.
  8. Abstract Wikipedia and Wikilambda will be editable by humans and by bots alike. But the people running the bots must be aware of their heightened responsibility to not overwhelm the community.


Architecture[edit]

The main components of the project are the following three:

  1. Constructors - definitions of Constructors and their slots, including what they mean and restrictions on the types for the slots and the return type of the Constructor (e.g. define a Constructor rank that takes in an item, an item type, the ranking as a number, what it is ranked by, and a local constraint)
  2. Content - abstract calls to Constructors including fillers for the slots (e.g. rank(SanFrancisco, city, 4, population, California))
  3. Renderers - functions that take Content and a language and return a text, resulting in the natural language representing the meaning of the Content (e.g. in the given example it results in "San Francisco is the fourth largest city by population in California.")

Components of Multilingual Wikipedia.png

There are four main possibilities on where to implement the three different main components:

  1. Constructors, Content, and Renderers are all implemented in Wikidata.
  2. Constructors and Renderers are implemented in Wikilambda (a new project), and the Content in Wikidata, next to the corresponding Item.
  3. Constructors, Content, and Renderers are all implemented in Wikilambda (a new project).
  4. Constructors and Content are implemented in Wikidata, and the Renderers in the local Wikipedias.

Solution 4 has the disadvantage that many functions can be shared between the different languages, and by moving the Renderers and functions to the local Wikipedias we forfeit that possibility. Also, by relegating the Renderers to the local Wikipedias, we miss out on the potential that an independent catalog of functions could achieve.

We think it is advantageous for communication and community building to introduce a new project, Wikilambda, for a new form of knowledge assets, functions, which include Renderers. This would speak for Solution 2 and 3.

Solution 3 requires us to create a new place for every possible Wikipedia article in Wikilambda. Given that a natural place for this already exists with the Items in Wikidata, it would be more convenient to use that and store the Content together with the Items in Wikidata.

Because of these reasons, we favor Solution 2 and assume it for the rest of the proposal If we switch to another, the project plan can be easily accommodated (besides for Solution 4, which would need quite some rewriting). Note that solution 2 requires the agreement of the Wikidata community to proceed. If they disagree, Solution 3 is likely the next closest option.

The proposed architecture for the multilingual Wikipedia looks as follows. Wikipedia calls the Content which is stored in Wikidata next to the Items. We call this extension of Wikidata Abstract Wikipedia. Note that this is merely a name for the development project, and that this is not expected to be a name that sticks around - there won’t be a new Wikiproject of that name. With a call to the Renderers in Wikilambda, the Content gets translated into natural language text. The Renderers rely on the other Functions, Types, and Constructors in Wikilambda. Wikilambda also can call out to the lexicographic knowledge in the Lexemes in Wikidata, to be used in translating the Content to text. Wikilambda will be a new Wikimedia project on par with Commons, Wikidata, or Wikisource.

Multlingual Wikipedia architecture.png

(The components named in italics are to be added by this proposal, the components in bold already exist. Top level boxes are Wikimedia projects, inner boxes are parts of the given Wikimedia projects.)


Components[edit]

We need to extend the Wikimedia projects in three places:

  1. in the local Wikipedias and other client projects using the new capabilities offered
  2. in Wikidata for creating Content (Abstract Wikipedia), and
  3. In a new project, Wikilambda, aimed to create a library of functions.

Extensions to local Wikipedias[edit]

Each local Wikipedia can choose, as per the local community, between one of the following three options:

  1. Implicit integration with Abstract Wikipedia
  2. Explicit integration with Abstract Wikipedia
  3. No integration with Abstract Wikipedia

The extension for local Wikipedias has the following functionalities: one new special page, two new features, and three new magic words.

F1: new Special Page: Abstract[edit]

A new Special page will be available on each local Wikipedia, that is used with a Q ID or the local article name and an optional language (which defaults to the language of the local Wikipedia). Example Special page URLs look like the following:

https://en.wikipedia.org/wiki/Special:Abstract/Q62
https://en.wikipedia.org/wiki/Special:Abstract/Q62/de
https://en.wikipedia.org/wiki/Special:Abstract/San_Francisco
https://en.wikipedia.org/wiki/Special:Abstract/San_Francisco/de

If the special page is called without parameters, then a form is displayed that allows for selecting a Q ID and a language (pre-filled to the local language).

The special page displays the Content from the selected Q-ID or the Q-ID sitelinked to the respective article rendered in the selected language.

F2: Explicit article creation[edit]

If the local Wikipedia chooses to go for the option of integrating with Abstract Wikipedia through explicit article creation, this is how they do it.

The contributor goes to an Item on Wikidata that does not have a sitelink in the target local Wikipedia yet. They add a sitelink to a page that does not exist yet. This way they specify the name of the article. For example, if Q62 in English would not have an article yet, and thus also no Sitelink, they may add the sitelink San_Francisco for en.wikipedia.

On the local Wikipedia, this creates a virtual article in the main namespace. That article has the same content as the special page described above, but it is to be found under the usual URL, i.e.

https://en.wikipedia.org/wiki/San_Francisco

Links to that article, using the newly specified name, look just like any other links, i.e. a link to [[San Francisco]] will point to the virtual article, be blue, etc. Such articles are indexed for search in the given Wikipedia and for external search too.

If a user clicks on editing the article, they can choose to either go to Wikidata and edit the abstract Content (preferred), or start a new article in the local language from scratch, or materialize the current translation as text and start editing that locally.

If an existing local article with a sitelink is deleted, a virtual article is automatically created (since we know the name and can retain the links).

In order to delete a virtual article, the sitelink in Wikidata needs to be deleted.

All changes to the local Wikipedia have to be done explicitly, which is why we call this the explicit article creation option. We expect to make this the default option for the local Wikipedias, unless they choose either implicit article creation or no integration.

See also the discussion on the integration here.

F3: Implicit article creation[edit]

If a local Wikipedia opts in to the Implicit article creation from Wikidata, then the result of calling the Abstract special page on all Wikidata Items that do not have a sitelink to the given Wikidata but would Render content in the given language, is indexed as if it were in the Main namespace, and made available in search as if it were in the Main namespace.

A new magic word is introduced to link to virtual articles from normal articles, see F6. This can be integrated invisibly into the visual editor.

This is by far the least work for the community to gain a lot of articles, and might be a good option for small communities.

F4: Links or tabs[edit]

Every article on a local Wikipedia that is connected to a Wikidata item receives a new link, either as a tab on the top or a link in the sidebar. That link displays the Content for the connected Wikidata item rendered in the local language. Virtual articles don’t have this tab, but their Edit button links directly to editing the Content in Abstract Wikipedia.

F5: New Magic Word: ABSTRACT_WIKIPEDIA[edit]

The magic word is replaced with the wikitext resulting from Rendering the Content on the Wikidata item that is connected to this page through sitelinks.

The magic word can be used with two optional parameters, one being a Q ID, the other a language. If no Q ID is given, the Q ID defaults to the Item this page is linked to per Sitelink. If no language is given, the language defaults to the language of the given wiki.

Example calls:

{{ABSTRACT_WIKIPEDIA}}
{{ABSTRACT_WIKIPEDIA:Q62}}
{{ABSTRACT_WIKIPEDIA:Q62/de}}

If no Q ID is given or chosen by default, an error message appears.

Later this will allow to select named sections from the Content.

Wikipedias that choose to have no integration to Abstract Wikipedia can still use this new magic word.

Note that the introduction of a new magic word is a preliminary plan. Task 2.3 will investigate whether we can achieve their functionalities without doing so.

F6: New Magic Word: LINK_TO_Q[edit]

This magic word turns into a link to either the local article that is sitelinked to the given Q ID or, if none exists, to the Abstract Special page with the given Q ID. This allows to write articles with links to virtual articles, which get replaced automatically once local content is created.

Example call:

{{LINK_TO_Q:Q62}}

will result in

[[San Francisco]]

if the article exists, otherwise in

[[Special:Abstract/Q62|San Francisco]]

Note that the introduction of a new magic word is a preliminary plan. Task 2.3 will investigate whether we can achieve their functionalities without doing so.

F7: New Magic Word: LAMBDA[edit]

This calls a function specified in Wikidata, together with its parameters, and renders the output on the page.

For example, the following call:

{{LAMBDA:capitalize(“san francisco”)}}

will result in ”San Francisco” being outputted on the page (assuming that there is a function that has the local key add with the expected definition and implementation). It uses the language of the local wiki to parse the call.

Consider also the option to call a specific version of a function in order to reduce breakages downstream.

Note that the introduction of a new magic word is a preliminary plan. Task 2.3 will investigate whether we can achieve their functionalities without doing so.

Extensions to Wikidata[edit]

We add a new auxiliary namespace to the main namespace of Wikidata. I.e. every item page of the form www.wikidata.org/wiki/Q62 will also receive an accompanying Content page www.wikidata.org/wiki/Content:Q62. That page contains the abstract, language-independent Content, and allows its editing and maintenance.

Additional Special pages might be needed. This will be extended in the second part of the project. It requires the agreement of the Wikidata community that the project will be used for storing the abstract content, and another will be chosen if they disagree.

F8: New Content namespace[edit]

New namespace with a lot of complex interactive editing features. Provides UX to create and maintain Contents, as well as features to evaluate the Contents (e.g. display how much of it is being displayed per language, etc.) This is mostly a subset of the functionality of the F9 Function namespace.

F9: New Content data type[edit]

A new datatype that contains a (short) Content. The main use case is for the Descriptions in Items and for Glosses in the Senses of Lexemes.

F10: Index and use Descriptions in Items and Glosses in Senses[edit]

Index and surface the linearizations of Descriptions in Items and Glosses in Senses, and also make sure that for Descriptions in Items there are no duplicate Label / Description pairs. Allow for overwriting these both by manual edits.

Extensions to other Wikimedia projects[edit]

Other Wikimedia projects will also receive F7 LAMBDA and F5 ABSTRACT_WIKIPEDIA magic words, but none of the other functionalities, as these seem not particularly useful for them. This may change based on requests from the given communities.

Extensions for Wikilambda[edit]

Wikilambda is a new Wikimedia project on a new domain. The main namespace of Wikilambda will be the novel Function namespace. The rest of Wikilambda will be a traditional Wikimedia wiki.

F11: New Function namespace[edit]

Allowing for the storage of functions, types, interfaces, values, tests, etc. There is a single namespace that contains constants (such as types or single values), function interfaces, function implementations, and thus also Constructors and Renderers. The entities in this namespace are named by Z-IDs, similar to Q-IDs of Wikidata items, but starting with a Z and followed by a number.

There are many different types of entities in the Z namespace. These include types and other constants (which are basically functions of zero arity), as well as classical functions with a positive arity.

Contributors can create new types of functions within the function namespace and then use these.

Functions can have arguments. Functions with their arguments given can be executed and result in a value whose type is given by the function definition.

The Function namespace is complex, and will have very different views depending on the type of the function, i.e. for interfaces, implementations, tests, types, values, etc. there will be different UX on top of them, although they are internally all stored as Z-Objects. Eventually, the different views are all generated by functions in Wikilambda.

It will be possible to freeze and thaw entities in the Function namespace. This is similar to a protected page, but only restricts the editing of the value part of the entity, not the label, description, etc.

F12: New Special pages and API modules[edit]

New Special pages and API modules will be created to support the new function namespace. This will include, in particular, a special page and an API module that allows to evaluate functions with function parameters given. Besides that it will include numerous special pages and APIs that will support the maintenance of the content (such as searches by number and types of parameters, pages with statistics of how often certain implementations are called, test pages, etc.). The goal is to implement as many as possible of these inside Wikilambda.


Tasks[edit]

Development happens in two main parts. Part P1 is about getting Wikilambda up and running, and sufficiently developed to support the Content creation required for Part P2. Part P2 is then about setting up the creation of Content within Wikidata, and allowing the Wikipedias to access that Content. After that, there will be ongoing development to improve both Wikilambda and Abstract Wikipedia. This ongoing development is not covered by this plan. Note that all timing depends on how much headcount we actually work with.

In the first year, we will work exclusively on Part P1. Part P2 starts with the second year and adds additional headcount to the project. Parts P1 and P2 will then be developed in parallel for the next 18 months. Depending on the outcome and success of the project, the staffing of the further development has to be decided around month 24, and reassessed regularly from then on.

This plan covers the initial 30 months of development. The main deliverables after that time will be:

  1. Wikilambda, a new WMF project and wiki, with its own community, and its own mission, aiming to provide a catalog of functions and allow everyone to share in that catalog, thus empowering people by democratizing access to computational knowledge
  2. A cross-wiki repository to share templates and modules between the WMF projects, which is a long-standing wish by the communities. This will be part of Wikilambda.
  3. Abstract Wikipedia, a cross-wiki project that allows to define Content in Wikidata, independently of a natural language, and that is integrated in several local Wikipedias, considerably increasing the amount of knowledge speakers of currently underserved languages can share in

Part P1: Wikilambda[edit]

Task P1.1: Project initialization[edit]

We will set up Wikipages on Meta, bug components, mailing lists, chat rooms, an advisory board, and other relevant means of discussion with the wider community. We will start the discussion on and decide on the names for Wikilambda and Abstract Wikipedia, and hold contests for the logos, organize the creation of a code of conduct, and the necessary steps for a healthy community.

We also need to kick off the community process of defining the license choice for the different parts of the project: abstract Content in Wikidata, the functions and other entities in Wikilambda, as well as the legal status of the generated text to be displayed in Wikipedia. This needs to be finished before Task P1.9. This decision will need input from legal counsel.

Task P1.2: Initial development[edit]

The first development step is to create the Wikilambda wiki with a Function namespace that allows the storage and editing of Z-Objects (more constrained JSON objects - in fact we may start with the existing JSON extensions and build on that). The initial milestone aims to have:

  • The initial types: unicode string, positive integer up to N, boolean, list, pair, language, monolingual text, multilingual text, type, function, builtin implementation, and error.
  • An initial set of constants for the types.
  • An initial set of functions: if, head, tail, is_zero, successor, predecessor, abstract, reify, equal_string, nand, constructors, probably a few more functions, each with a built-in implementation.

This allows to start developing the frontend UX for function calls, and create a first set of tools and interfaces to display the different types of Z Objects, but also generic Z Objects. This also includes a first evaluator that is running on Wikimedia servers.

Note that the initial implementations of the views, editing interfaces, and validators are likely to be thrown away gradually after P1.12 once all of these are becoming internalized. To internalize some code means to move it away from the core and move it into userland, i.e. to reimplement them in Wikilambda and call them from there.

Task P1.3: Set up testing infrastructure[edit]

Wikilambda will require several test systems. One to test the core, one to test the Web UI, one to test the Wikilambda content itself. This is an ongoing task and needs to be integrated with version control.

Task P1.4: Launch public test system[edit]

We will set up a publicly visible and editable test system that runs the bleeding edge of the code (at least one deploy per working day). We will invite the community to come in and break stuff. We may also use this system for continuous integration tests.

Task P1.5: Server-based evaluator[edit]

Whereas the initial development has created a simple evaluator working only with the built-ins, and thus having very predictable behavior, the upcoming P1.6 function composition task will require us to rethink the evaluator. The first evaluator will run on Wikimedia infrastructure, and needs monitoring and throttling abilities, and potentially also the possibility to allocate users different amounts of compute resources depending on whether they are logged-in or not.

Task P1.6: Function composition[edit]

We allow for creating new function interfaces and a novel type of implementation, which are composed function calls. E.g. it allows to implement

add(x, y)

as

if(is_zero(y), x, add(successor(x), predecessor(y))

and the system can execute this. It also allows for multiple implementations.

Task P1.7: Freeze and thaw entities[edit]

A level of protection that allows to edit the metadata of an entity (name, description, etc.), but not the actual value. This functionality might also be useful for Wikidata. A more elaborate versioning proposal is suggested here.

Task P1.8: Launch beta Wikilambda[edit]

The beta system runs the next iteration of the code that will go on Wikilambda with the next deployment cycle, for testing purposes.

Task P1.9: Launch Wikilambda[edit]

Pass security review. Set up a new Wikimedia project. Move some of the wikipages from Meta to Wikillambda.

Task P1.10: Test type[edit]

Introduce a new type for writing tests for functions. This is done by specifying input values and a function that checks the output. Besides introducing the Test type, Functions and Implementations als have to use the Tests, and integrate them into Implementation development and Interface views.

Task P1.11: Special page to write function calls[edit]

We need a new Special page that allows and supports to write Function calls, with basic syntax checking, autocompletion, documentation, etc. A subset of this functionality will also be integrated on the pages of individual Functions and Implementations to run them with more complex values. This Special page relies on the initial simple API to evaluate Function calls.

Task P1.12: JavaScript-based implementations[edit]

We allow for a new type of Implementations, which are implementations written in JavaScript. This requires to translate values to JavaScript and back to Z-Objects. This also requires to think about security, through code analysis and sandboxing, initially enriched by manual reviews and P1.7 freezing. If the O1 lambda-calculus has not yet succeeded or was skipped, we can also reach the self-hosting for Wikilambda through this task, which would allow to internalize most of the code.

Task P1.13: Access functions[edit]

Add F7 the Lambda magic word to the Wikimedia projects which can encode function calls to Wikilambda and integrate the result in the output of the given Wikimedia project. This will, in fact, create a centralized templating system as people realize that they can now reimplement templates in Wikilambda and then call them from their local Wikis. This will be preceded by an analysis of existing solutions such as TemplateData and the Lua calls.

This might lead to requests from the communities to enable the MediaWiki template language and Lua (see P1.15) as a programming language within Wikilambda. This will likely lead to requests for an improved Watchlist solution, similar to what Wikidata did, and to MediaWiki Template-based implementations, and other requests from the community. In order to meet these tasks, an additional person might be helpful to answer the requests from the community. Otherwise P2 might start up to a quarter later. This person is already listed in the development team above.

Task P1.14: Create new types[edit]

We allow for the creation of new types. This means we should be able to edit and create new type definitions, and internalize all code to handle values of a type within Wikilambda. I.e. we will need code for validating values, construct them, visualize them in several environments, etc.. We also should internalize these for all the existing types.

Task P1.15: Lua-based implementations[edit]

We add Lua as a programming language that is supported for implementations. The importance of Lua is due to its wide use in the MediaWiki projects. Also, if it has not already happened, the value transformation from Wikilambda to a programming language should be internalized at this point (and also done for the JavaScript implementations, which likely will be using built-ins at this point).

Task P1.16: Non-functional interfaces[edit]

Whereas Wikilambda is built on purely functional implementations, there are some interfaces that are naively not functional, for example random numbers, current time, autoincrements, or many REST calls. We will figure out how to integrate these with Wikilambda.

Task P1.17: REST calls[edit]

We will provide builtins to call out to REST interfaces on the Web and ingest the result. This would preferably rely on P1.16. Note that calling arbitrary REST interfaces has security challenges. These need to be taken into account in a proper design.

Task P1.18: Accessing Wikidata and other WMF projects[edit]

We will provide Functions to access Wikidata Items and Lexemes, and also other content from Wikimedia projects. This will preferably rely on P1.17, but in case that didn’t succeed yet at this point, a builtin will unblock this capability.

Task P1.19: Monolingual generation[edit]

The development of these functions happen entirely on-wiki. This includes tables and records to represent grammatical entities such as nouns, verbs, noun phrases, etc., as well as Functions to work with them. This includes implementing context so that we can generate anaphers as needed. This allows for a concrete generation of natural language, i.e. not an abstract one yet.

Part P2: Abstract Wikipedia[edit]

The tasks in this part will start after one year of development. Not all tasks from Part P1 are expected to be finished by then, as, in fact, Parts P1 and P2 will continue to be developed in parallel. Merely the parts that are necessary for Part P2 to start need to be finished for P2 to start.

Task P2.1: Constructors and Renderers[edit]

Here we introduce the abstract interfaces to the concrete generators developed in P1.19. This leads to the initial development of Constructors and the Render function. After this task, the community should be able to create new Constructors and extend Renderers to support them.

Task P2.2: Conditional rendering[edit]

It will rarely be the case that a Renderer will be able to render the Content fully. We will need to support graceful degradation: if some Content fails to render, but other still renders, we should show that part that rendered. But sometimes it is narratively necessary to render certain Content only if other Content will definitely be rendered. In this task we will implement the support for such conditional rendering, that will allow smaller communities to grow their Wikipedias safely.

Task P2.3: Abstract Wikipedia[edit]

Create a new namespace in Wikidata and allow Content to be created and maintained there. Reuse UI elements and adapt them for Content creation. The UI work will be preceded by Design research work which can start prior to the launch of Part P2. Some important thoughts on this design are here. This task will also decide whether we need new magic words (F5, F6, and F7) or can avoid their introduction.

Task P2.4: Mobile UI[edit]

The creation and editing of the Content will be the most frequent task in the creation of a multilingual Wikipedia. Therefore we want to ensure that this task has a pleasant and accessible user experience. We want to dedicate an explicit task to support the creation and maintenance of Content in a mobile interface. The assumption is that we can create an interface that allows for a better experience than editing wikitext.

Task P2.5: Integrate content into the Wikipedias[edit]

Enable the Abstract Wikipedia magic word. Then allow for the explicit article creation, and finally the implicit article creation (F1, F2, F3, F4, F5, F6).

Task P2.6: Regular inflection[edit]

Wikidata’s Lexemes contain the inflected Forms of a Lexeme. These Forms are often regular. We will create a solution that generates regular inflections through Wikilambda and will discuss with the community how to integrate that with the existing Lexemes.

Task P2.7: Basic Renderer for English[edit]

We assume that the initial creation of Renderers will be difficult. Given the status of English as a widely used language in the community, we will use English as a first language to demonstrate the creation of a Renderer, and document it well. We will integrate the help from the community. This also includes functionality to show references.

Task P2.8: Basic Renderer for a second language[edit]

Based on the community feedback, interests, and expertise of the linguists working on the team, we will select a second large language for which we will create the basic Renderer together with the community. It would be interesting to choose a language where the community on the local Wikipedia has already confirmed their interest to integrate Abstract Wikipedia.

Task P2.9: Renderer for a language from another family[edit]

Since it is likely that the language in P2.8 will be an Indo-European language, we will also create a basic Renderer together with the community for a language from a different language family. The choice of this language will be done based on the expertise available to the team and the interests of the community. It would be interesting to choose a language where the community on the local Wikipedia has already confirmed their interest to integrate Abstract Wikipedia.

Task P2.10: Renderer for an underserved language[edit]

Since it is likely that the languages in P2.8 and P2.9 will be languages which are already well-served with active and large Wikipedia communities, we will also select an underserved language, a language that currently has a large number of potential readers but only a small community and little content. The choice of this language will be done based on the expertise available to the team and the interests of the community. Here it is crucial to select a language where the community on the local Wikipedia has already committed their support to integrate Abstract Wikipedia.

Task P2.11: Abstract Wikidata Descriptions[edit]

Wikidata Descriptions seem particularly amenable to be created through Wikilambda. They are often just short noun phrases. In this task we support the storage and maintenance of Abstract Descriptions in Wikidata, and their generation for Wikidata. We also should ensure that the result of this leads to unique combinations of labels and descriptions.

Task P2.12: Abstract Glosses[edit]

Wikidata Lexemes have Senses. Senses are captured by Glosses. Glosses are available per language, which means that they are usually only available in a few languages. In order to support a truly multilingual dictionary, we are suggesting to create Abstract Glosses. Although this sounds like it should be much easier than creating full fledged Wikipedia articles, we think that this might be a much harder task due to the nature of Glosses.

Task P2.13: Support more natural languages[edit]

Support other language communities in the creation of Renderers, with a focus on underserved languages.

Task P2.14: Template-generated content[edit]

Some Wikipedias currently contain a lot of template-generated content. Identify this content and discuss with the local Wikipedias whether they want to replace it with a solution based on Wikilambda, where the template is in Wikilambda and the content given in the local Wikipedia or in Abstract Wikipedia. This will lead to more sustainable and maintainable solutions that do not require to rely on a single contributor. Note that this does not have to be multilingual, and might be much simpler than going through full abstraction.

Task P2.15: Casual comments[edit]

Allow casual contributors to make comments on the rendered text and create mechanisms to capture those comments and funnel them back to a triage mechanism that allows to direct them to either the content or the renderers. It is important to not lose comments by casual contributors. Ideally, we would allow them to explicitly overwrite a part of the rendered output and consider this a change request, and then we will have more involved contributors working on transforming the intent of the casual contributor to the corresponding changes in the system.

Task P2.16: Quick article name generation[edit]

The general public comes to Wikipedia mostly by typing the names of the things they are looking for in their language into common search engines. This means that Wikidata items will need labels translated into a language in order to be able to use implicit article creation. This can probably be achieved by translating millions of Wikidata labels. Sometimes it can be done by bots or AI, but this is not totally reliable and scalable, so it has to involve humans.

The current tools for massive crowd-sourced translation of Wikidata labels are not up to the task. There are two main ways to do it: editing labels in Wikidata itself, which is fine for adding maybe a dozen of labels, but quickly gets tiring, and using Tabernacle, which appears to be more oriented at massive batch translations, but is too complicated to actually use for most people.

This task is to develop a massive and integrated label-translation tool with an easy, modern frontend, that can be used by lots of people.

Non-core tasks[edit]

There is a whole set of further optional tasks. Ideally these would be picked up by external communities and developed as Open Source outside the initial development team, but some of these might need to be kickstarted and even fully developed by the core team.

Task O1: Lambda calculus[edit]

It is possible to entirely self-host Wikilambda without relying on builtins or implementations in other programming languages, by implementing a Lambda calculus in Wikilambda (this is where the name proposal comes from). This can be helpful to allow evaluation without any language support, and so easier start up the development of evaluators.

Task O2: CLI in a terminal[edit]

Many developers enjoy using a command line interface to access a system such as Wikilambda. We should provide one, with the usual features such as autocompletion, history, shell integration, etc.

Task O3: UIs for creating, debugging and tracing functions[edit]

Wikilambda’s goal is to allow the quick understanding and development of the functions in Wikilambda. Given the functional approach, it should be possible to create a user experience that allows for partial evaluation, unfolding, debugging, and tracing of a function call.

Task O4: Improve evaluator efficiency[edit]

There are many ways to improve the efficiency of evaluators, and thus reduce usage of resources, particularly caching, or the proper selection of an evaluation strategy. We should spend some time on doing so for evaluators in general and note down the results, so that different evaluators can use this knowledge, but also make sure that the evaluators maintained by the core team use most of the best practises.

Task O5: Web of Trust for implementations[edit]

In order to relax the conditions for implementations in programming languages, we could introduce a Web of Trust based solution, that allows contributors to review existing implementations and mark their approval explicitly, and also to mark other contributors as trustworthy. Then these approvals could be taken into account when choosing or preparing an evaluation strategy.

Task O6: Python-based implementations[edit]

Python is a widely used programming language, particularly for learners and in some domains such as machine learning. Supporting Python can open a rich ecosystem to Wikilambda.

Task O7: Implementations in other languages[edit]

We will make an effort to call out to other programming language communities to integrate them into Wikilambda, and support them. Candidates for implementations are Web Assembler, PHP, Rust, C/C++, R, Swift, Go, and others, but it depends on the interest of the core team and the external communities to create and support these interfaces.

Task O8: Web-based REPL[edit]

A web-based REPL can bring the advantages of the O2 command line interface to the Web, without requiring to install a CLI in a local environment, which is sometimes not possible.

Task O9: Extend API with Parser and Linearizer[edit]

There might be different parsers and linearizers using Wikilambda. The Wikilambda API can be easier to use if the caller could explicitly select these, instead of wrapping them manually, which would allow the usage of Wikilambda with different surface dialects.

Task O10: Support talk pages[edit]

In order to support discussions on talk pages of Wikilambda, develop and integrate a mechanism that allows for (initially) simple discussions, and slowly raising their complexity based on the need of the communities.

Task O11: Legal text creation[edit]

An interesting application of Wikilambda is for the creation of legal text in a modular manner and with different levels (legalese vs human-readable), similar to the different levels of description for the different Creative Commons licenses.

Task O12: Health related text creation[edit]

An interesting application of Wikilambda is for the creation of health-related texts for different reading levels. This should be driven by WikiProject Medicine and their successful work, which could reach many more people through cooperation with Wikilambda.

Task O13: NPM library[edit]

We will create a Wikidata library for NPM that allows the simple usage of Functions from Wikilambda in a JavaScript program. The same syntax should be used to allow JavaScript implementations in Wikilambda to access other Wikilambda functions. Note that this can be done by virtue of calling to a Wikilambda evaluator, or by compiling the required functions into the given code base.

Task O14: Python library[edit]

We will create a Python library that allows the simple usage of Functions from Wikilambda in a Python script. The same syntax should be used to allow Python implementations in Wikilambda to access other Wikilambda functions. Note that this can be done by virtue of calling to a Wikilambda evaluator, or by compiling the required functions into the given code base.

Task O15: Libraries for other programming languages[edit]

We will call out to the communities of several programming languages to help us with the creation of libraries that allow the simple call of Wikilambda functions from programs in their language. Note that this can be done by virtue of calling to a Wikilambda evaluator, or by compiling the required functions into the given code base.

Task O16: Browser-based evaluator[edit]

One of the advantages of Wikidata is that the actual evaluation of a function call can happen in different evaluators. The main evaluator for Abstract Wikipedia will be server-based and run by the Wikimedia Foundation, but in order to reduce the compute load, we should also provide evaluators running in the user’s client (probably in a Worker thread).

Task O17: Jupyter- and/or PAWS-based evaluator[edit]

One interesting evaluator is from a Jupyter or PAWS notebook, and thus allowing the usual advantages of such notebooks but integrating also the benefits from Wikilambda.

Task O18: App-based evaluator[edit]

On evaluator should be running natively on Android or iOS devices, and thus allow the user to use the considerable computing power in their hand.

Task O19: P2P-based evaluator[edit]

Many of the evaluators could be linking together and allow each other to use dormant compute resources in their network. This may or may not require shields between the participating nodes that ensure the privacy of individual computes.

Task O20: Cloud-based evaluator[edit]

One obvious place to get compute resources is by allowing to use a cloud provider. Whereas it would be possible to simply run the server-based evaluator on a cloud-based infrastructure, it is likely to be beneficial for the cloud providers to provide a thinner interface to a more custom-tailored evaluator.

Task O21: Stream type[edit]

Add support for a type for streaming data, both as an input as well as an output. Stream types means for example the recent changes stream on a Wikimedia wiki.

Task O22: Binary type[edit]

Add support for binary files, both for input and output.

Task O23: Integration with Commons media files[edit]

Allow direct access to files on Commons. Enable workflows with Commons that require less deployment machinery than currently needed. Requires O22.

Task O24: Integrate with Machine Learning[edit]

Develop a few example integrations with Machine Learning solutions, e.g. for NLP tasks or for work on image or video e.g. using classifiers. This requires how and where to store models, possibly also how to train them, and how to access them.

Task O25: Integrate into IDEs[edit]

Reach out to communities developing IDEs and support them in integrating with Wikilambda, using type hints, documentation, completion, and many of the other convenient and crucial features of modern IDEs.

Task O26: Create simple apps or Websites[edit]

Develop a system to allow the easy creation and deployment of apps or Websites relying on Wikilambda.

Task O27: Display open tasks for contributors[edit]

Abstract Wikipedia will require one Renderer for every Constructor for every language. It will be helpful if contributors could get some guidance to which Renderer to implement next, as this is often not trivially visible. Just counting how often a Constructor appears could be a first approximation, but if some Constructors are used more often in the lead, or in parts of the text that block other text from appearing, or in articles that are more read than others, this approximation might be off. In this task we create a form of dashboard that allows users to choose a language (and maybe a domain, such as sports, geography, history, etc., and maybe a filter for the complexity of the renderer that is expected) and then provide them with a list of unrendered Constructors ranked by the impact an implementation would have.

We could also allow contributors to sign up for a regular message telling them how much impact (in terms of views and created content) they had, based on the same framework needed for the dashboard.

This is comparable to seeing the status of translations of different projects on translatewiki or the views on topics, organizations or authors in Scholia. For each project, it shows what % of the strings in it were translated and what % need update, and a volunteer translator can choose: get something from 98% to 100%, get something from 40% to 60%, get something from 0% to 10%, etc.

Task O28: Active view[edit]

Whereas the default view of Rendered content would look much like static text, there should also be a more active view that invites contribution, based on existing Content that failed to render due to missing Renderers. In the simplest case this can be the creation of a Lexeme in Wikidata and connecting to the right Lexeme. In more complex cases that could be writing a Renderer, or offering example sentences as text, maybe using the Casual contribution path described in P2.15. This would provide an interesting funnel to turn more readers in contributors.

There are product and design decisions on how and where the active view should be used and whether it should be the default view, or whether it should be only switched on after an invitation, etc. There could also be mode where contributors can go from article to article and fill out missing pieces, similar to the more abstract way in O27.

It would probably be really useful that ensure that the active way and the contribution path it leads to work on mobile devices as well.