SWIB 19

Draft  SWIB 19 submission

Wikidata/Wikipedia is a project supporting +300 languages and in Wikidata we have defined more than xxx external properties --> we get an echo system of loosely coupled actors inside Wikidata but  also between Wikidata and external sources  that Wikidata needs a way to cooperate with and also create TRUST between each other.

This session is about my experience coming from beeing a Scrum Master at a big bank in Sweden where traceability is very very very important to now being a voluntaire on a Wiki platform called Wikidata were we try to add value to the LOD community.

Last year at SWIB18 Wikidata was presented by Karen Smith-Yoshimura as one of the fastest growing members video

What is needed to do LOD between loosely coupled networks

To get this working we need to define what is best practices:

• define a change process that is communicated
• have traceability on open tasks
• open public backlogs
• have a roadmap that you can easily follow/subscribe to --> the benefit will come

The key components I feel is that we need to work with TRUST plus  LOD knowledge

and if we get more TRUST between participators then we will do more things see FutureLearn "Introduction to strategic doing"

                  



Anti-patterns:

1. we are a national library we need no sources 
2. doorkeepers on Facebook
3. being afraid of saying No
4. communicating without traceability 
1. is a big waste of time for both parties

Good members in the Linked Data Landscape 

Wikidata

• Traceability using Phabricator see Project space Wikidata
• Road map 2019 with links to Tasks that you can subscribe to
• Fast transparent error reporting using Phabricator

Google

Very fast moving but difficult to understand what they do. Example (see tweet) it looks like they read the Wikidata change stream. 

I added in Wikidata coordinates for a building Älvrosgården at an open air museum --> 20 minutes later I could see that they had read the Wikidata stream and

• connected pictures from that location
• added reviews
• changed the search result
  
  

Anti-patterns seen 

Swedish LIBRIS XL:

1. having doorkeepers or people with not the right skill matrix answering question. 
1. I asked about a year ago about how the new LIBRIS XL should work with VIAF and still no answers plus we start to see problems that could indicate VIAF mix ld and new IDs
   
2. They have no public backlog, users complaining in discussion groups of the bad interaction etc.  its impossible to understand what they release
   (odd thing the source code is managed in public on GITHUB --> they have tools tat supports traceability but has decided not to use them )

Swedish National Heritage RAÄ

has no well defined way of communicating --> twitter, Facebook is used and no one knows of open issues, what was said last time we communicated

Swedish National Archive

has no well defined way of communicating --> twitter, Facebook is used and no one knows of open issues, what was said last time we communicated

lack of traceability etc....

How to move this forward

• Communicating
• Communicating
• Communicating and start set up public workboards is my suggestion. Example what I have use  Wikimedia Phabricator

See also:

• The rise of Wikidata as a linked data source August 6, 2018 - by Karen Smith-Yoshimura
• Me speaking at Linked Data Sweden 2019
• Video / presentation
  
  
• tweet
• Facebook group I manage about Genealogy and semantic Web (in Swedish)