As Wikidata now will support Shape Ex. One nice pattern I guess is to define what "Riksarkivet" has and see what WIkidata supports regarding Svenska Församlingar / Svenska socknar / Domsagor...
Description
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Open | None | T224088 Omstart 3 Riksarkivet | |||
Open | None | T224336 Define ShapeEx for Swedish parishes / judicial district for Riksarkivet and Wikidata |
Event Timeline
What do you mean by this? The RAÄ parish data(SOCH(SOCH does however use ShEx for other data)) has no need for ShEx for because the data is static and can only be edited after review by "experts" I think the same is true for Riksarkivets parish data.
The Wikidata community could however model ShEx rules for the parish Wikidata items but it has absolute nothing to do with RAÄ or Riksarkivet.
My mistake should be Riksarkivet....
In the Tora project Riksarkivet will have of Swedish Parishes. I hope that Riksarkivet
- will have a better more advanced approach than Wikidata
- when I spoke with them about a year ago they spoke about first documented evidence. The idea is that we learn from them and see how if we can implement that in Wikidata or if it is overkill etc.
- could help us check the data in Wikidata. We are seeing more and more odd edits and bots that change data link. The best approach I guess is using SPARQL federation and compare WD <-> TORA and maybe also ShapeEx can help. Lesson learned from Wikidata <-> Nobelprize.org is that comparing works excellent BUT Nobelprize has no sources/references proving the data they have in he facts so its nearly impossible to say if Wikidata or Nobelprize is right...
ShEx is the Wikidata "design" /Schema. By formalize it in a schema it will be much easier to communicate and compare with external sources..... I spoke with Olof at Riksarkivet about that todays "copy everything" pattern is not good
I hope ShEx could be some kind of "contract" that people inside Wikidata agree about and we validate our WD with and this also communicates our ambition level.
Maybe first documented evidence for a parish can be linked data that always will be fetched from Riksarkivet.... ==> we should have some kind of "cache objects" in Wikidata that always are fetched or maybe federation has good enough performance...
For the Nobelprize winners we fetch prize motivation from Nobelprize.org I guess by defining ShEx for Nobelprize winners inside WIkidata and at Nobelprize.... doing it with ShEx would make sense.....
Nobelprize has today a good design see below but using a standard makes more sense..
I guess it would make sense also to define a "Wikidata Swedish church" and compare it with a much more advanced "RAÄ church".... and a design would maybe also make sense when you start looking at data roundtrips and what parts should be "returned"
I think its things like that Digisam should push and get a GITHUB ShEx Sweden Culture that everyone can read and comment...
My feeling after Olof presentation last week was that he is moving in the right direction but most people in the audience didnt feel so interested.... and I can feel the project is not pushed as much it should.
- WD created the TORA ID P4820 2018 feb 8 and still nothing has happened,,,,
- the people at SKBL design a new website and dont link TORA instead use string see T218782: SKBL <-> KARP Svenskt kvinnobiografiskt lexikon
I'm fully aware of what ShEx is and how it's used.
How would ShEx help you compare data between datasets? That's what we have SPARQL for, ShEx allows you to validate and find data errors within a dataset.
Your TORA ID property has nothing to do with this and the claim you make in the property proposal that it's used by RAÄ is false.
RAÄ: Ok update
Maybe the concept is the same as Wikidata that Tora will just have the links!!
my feeling is that the National Archive would like that see pdf
RAÄ was mentioned as potential at Digikult 2016 Clas Tollin min 6:32
In https://toramaptest.riksarkivet.se/ we see a non working option Fornsök
I am not fully aware of the potential of ShEx someone indicated that ShEx could be the solution for saying that WD Object xxx Property yyy has a value that differs from what external source at the National Archives has which is something that fits into my vision I asked for in Berlin 2017 at Wikidata conf link at min 42:08