Page MenuHomePhabricator

Wikibase Bootstrap set, a year retrospective and discussion for the future
Closed, ResolvedPublic

Description

A year ago (T391815) for the Wikimedia-Hackathon-2025 was proposed to select a minimalist set of properties for blank Wikibase instances. The original goal was to solve the blank page syndrome for Wikibase newcomers having some basic properties, mapped with Wikidata, to start their experiments and reduce their learning curve. For this use, it could be ideal for the Wikibase Cloud data example preload. A first release has been published at https://gitlab.wikimedia.org/olea/wikibase-bootstrap/

Also, with the development of an OWL to Wikibase workflow at crm4wb we found the importance of the set as an interoperability standard for tools. But then things when bigger: when you try to productize Wikibase (like we try in VerySmallGLAM) you find you'll need to set more conventions. This happens for tools (like crm4wb) and to advanced configurations (like this fork of WBS for GLAM), so an extension set is being created attending the requirements appearing while developing those artifacts.

Why this will be important? When adding new tools to the Wikibase ecosystem, in many cases we are generalizing tools created and set according to specific Wikidata entities. At this moment, if you need to include those tools (say OpenRefine reconcilator, Wikiibase Manifest, SDC...) in your practical Wikibase implementation you should choose arbitrary configuration values. An specific goal for the extended set is to include standardize the configuration values for the most important tools in the ecosystem for reducing the entry barriers for more complex and powerful instances. Also, new features like Wikibase federation will need to set provenance practices (see T420777) and many other existing Wikibase features (say QualityConstrains —see T423246— or references) need preconfigurations to be easy to adopt.

Please come to this session and let's discuss present and future.


Raw session notes

Unconference session: Wikibase Bootstrap set, a year retrospective and discussion for the future (T423677)

It's sometimes hard to define notability criteria for certain items are not notable enough.
Q: Do yaou also have the situation that you want to model data in a different way, does this happen to you?
A: Yes, I've run into situations where we want to model a class, for ex. A transgender person as a class we a like to define it as a concept rather than a class. Currently, Wikidata doesn't include many values that are on Wikidata but I think it's because we have hundreds of values on [sora].

One of the largest publications in the US is called the '', which unfortunately if you want to access editions made before 2012 you can't, because they were not archived.

So I started archiving everything, there's an RSS collector where I put the RSS feeds for publications I'm interested in. So everyday I scrape a lot of LGBTQ publications and I add everything to our Wiki,

Q: Are you using a library to access Wikidata APIs?
A: I use a package called Wikibaseintegrator, it works on any wikibase (not just Wikidata). I like Wikiteams 3 and it has an automatic upload to the Internet Archive.

Date & time: Saturday, May 2nd, 2026 at 1445 am CEST

Note taker

Anonymous

Participants

  • Maybe 10 persons
  • ...

Relevant links

Notes

Maybe the first standard for Wikibase ecosystem.
Primarily a bunch of properties

Rationale

To fight he effect of the blank page block... You can be very experienced with Wikidata, However, I was frozen with the blank page block of wikibase, So I started thinking about the tooling case, of a set of properties that everybody needs to setup on any kind of instance

When you first start a wikibase it can be dificult
If you are a museum, and you have artificats and want countries of origin, you have to initially create that data to use, and this is tiring.

Currently this is only 8 properties, including descriptions and equivilent on Wikidata
These are essetnial
The real work should be done on your own ontology

My idea was that the first usage would be on wikibase cloud
As there is already a button to fill your instance with some sample data
I was going to wrtie a phabricator ticket
But thne "Wikidata item" property poses a problem, as it doesnt exist on wikidata to use as a mapping...

There is also an extended set

P9, inverse of
Needed in order to represent some onotologies accruatly (in the spirit of their source)

P10, depicts
If I want to use the extension StructuredDataForCommons
I need to hard code this property as part of the extension setup

If we want interoperability
We need to agree on some things
and consider more features than we currently do

If you use the wikibase reconciliator for open refine
it expexts to read some properties from a manifest (quality constraints etc)
You can do everything by hand,
but my personal goal is to ahve everything automatic

We want to make this as big as possible, to reduce time to setup and increase usefullness

I would like feedback
I would like more people to adopt the standard
The ore features that you includ in your wikibase, the more of these properties you will need

As an example usecase
If you live in a country where people want data located in their physical location
Where is book X in a set of library
Then using a defualt set (along with federated search) would relaly change the game, in terms of connecting instances together

Questions

I crated my first WB cloud instance a second go
I dont know where i should go to create my bootstrap
<points at the ttl2wb.py wikibase-bootstrap.ttl>

Martyn: Am I understanding your question differently to others
I understand now this .env file has the infomation for the repo to know which one is my wikibase and credentials etc.
You have to run the script

Can we turn it into a tool
Yes

We talked about the numbering of the properties on wikibases

Doing that in quickstaements would be like .....

Ollie: If hte ids are important, then it can only be done on a new wiki?
So you cant do this on an existing wikibase
So you have to do it on a fresh install
A: Yes
Ollie: Is that a deal breaker, or can we come up with a way of saying on my wikibase P123 is X P5676 i Y etc

Details

Other Assignee
Olea

Event Timeline

Notes from https://etherpad.wikimedia.org/p/2026-wikibase-bootstrap-hackathon

Date & time: Saturday, May 2nd, 2026 at 1445 am CEST

Note taker

Anonymous

Participants

Maybe 10 persons
...

Relevant links

Main reference: http://gitlab.wikimedia.org/olea/wikibase-bootstrap

Notes

Maybe the first standard for Wikibase ecosystem.
Primarily a bunch of properties

Rationale

To fight he effect of the blank page block... You can be very experienced with Wikidata, However, I was frozen with the blank page block of wikibase, So I started thinking about the tooling case, of a set of properties that everybody needs to setup on any kind of instance

When you first start a wikibase it can be dificult
If you are a museum, and you have artificats and want countries of origin, you have to initially create that data to use, and this is tiring.

Currently this is only 8 properties, including descriptions and equivilent on Wikidata
These are essetnial
The real work should be done on your own ontology

My idea was that the first usage would be on wikibase cloud
As there is already a button to fill your instance with some sample data
I was going to wrtie a phabricator ticket
But thne "Wikidata item" property poses a problem, as it doesnt exist on wikidata to use as a mapping...

There is also an extended set

P9, inverse of
Needed in order to represent some onotologies accruatly (in the spirit of their source)

P10, depicts
If I want to use the extension StructuredDataForCommons
I need to hard code this property as part of the extension setup

If we want interoperability
We need to agree on some things
and consider more features than we currently do

If you use the wikibase reconciliator for open refine
it expexts to read some properties from a manifest (quality constraints etc)
You can do everything by hand,
but my personal goal is to ahve everything automatic

We want to make this as big as possible, to reduce time to setup and increase usefullness

I would like feedback
I would like more people to adopt the standard
The ore features that you includ in your wikibase, the more of these properties you will need

As an example usecase
If you live in a country where people want data located in their physical location
Where is book X in a set of library
Then using a defualt set (along with federated search) would relaly change the game, in terms of connecting instances together

Questions

I crated my first WB cloud instance a second go
I dont know where i should go to create my bootstrap
<points at the ttl2wb.py wikibase-bootstrap.ttl>

Martyn: Am I understanding your question differently to others
I understand now this .env file has the infomation for the repo to know which one is my wikibase and credentials etc.
You have to run the script

Can we turn it into a tool
Yes

We talked about the numbering of the properties on wikibases

Doing that in quickstaements would be like .....

Ollie: If hte ids are important, then it can only be done on a new wiki?
So you cant do this on an existing wikibase
So you have to do it on a fresh install
A: Yes
Ollie: Is that a deal breaker, or can we come up with a way of saying on my wikibase P123 is X P5676 i Y etc