Hi,
My view is that, in order for there to be any point in cultural heritage bodies (museums, libraries, archives, historians) publishing their collections etc. as Linked Data, there needs to be a common Linked Data framework representing the historical space-time universe, which they can all quote. Current practice (such as the British Museum Linked Data offering) suggests that concepts such as people, places and events will otherwise be represented either by useless string values or by system-specific URLs which have no wider meaning.
As a result, I would like to explore the potential for Wikidata to act as this lingua franca for the cultural heritage community.
You'll see from my earlier messages to this list that I have been grappling with the SPARQL end-point. Initially I was confused by the interactive version of the Query Service [1], which differs in its response format from the similarly-URLed end-point and doesn't provide an RDF/XML response. I have now managed to set up Wikidata as a 'web termlist' service for artists, within the Modes software (see attached screenshot). (The data in the pop-up window is generated on the fly from the Wikidata RDF.)
At this point, I have the following questions:
1. what level of stability is planned as regards Wikidata identifiers/URLs? Can I treat the full URL (e.g.[3]) as persistent, or can I only rely on the core Wikidata identifier (e.g. [4]) remaining unchanged into the indefinite future? (Can I even rely on that?)
2. what is the policy on inclusivity? Do entities need to be 'notable' in some sense to be accepted into Wikidata? (I'm imagining a research body wanting to offer very precise place or event data, or someone with the ambition to include in Wikidata details of any person who ever lived.)
3. is there a template for each entity type (e.g. person, place, event) which guarantees that a query for certain properties will at least identify entities of the desired type? (My artist termlist query includes a test '$s ps:P31 wd:Q5' which picks out humans: I'm not clear how I would do the same for events or places.)
Thanks,
Richard
[1] https://query.wikidata.org/#%7BSPARQL%7D [2] https://query.wikidata.org/sparql?query=%7BSPARQL%7D [3] http://www.wikidata.org/entity/Q2591521 [4] Q2591521
On Sun, Jun 19, 2016 at 1:27 PM, Richard Light richard@light.demon.co.uk wrote:
Hi,
My view is that, in order for there to be any point in cultural heritage bodies (museums, libraries, archives, historians) publishing their collections etc. as Linked Data, there needs to be a common Linked Data framework representing the historical space-time universe, which they can all quote. Current practice (such as the British Museum Linked Data offering) suggests that concepts such as people, places and events will otherwise be represented either by useless string values or by system-specific URLs which have no wider meaning.
As a result, I would like to explore the potential for Wikidata to act as this lingua franca for the cultural heritage community.
You'll see from my earlier messages to this list that I have been grappling with the SPARQL end-point. Initially I was confused by the interactive version of the Query Service [1], which differs in its response format from the similarly-URLed end-point and doesn't provide an RDF/XML response. I have now managed to set up Wikidata as a 'web termlist' service for artists, within the Modes software (see attached screenshot). (The data in the pop-up window is generated on the fly from the Wikidata RDF.)
At this point, I have the following questions:
- what level of stability is planned as regards Wikidata identifiers/URLs?
Can I treat the full URL (e.g.[3]) as persistent, or can I only rely on the core Wikidata identifier (e.g. [4]) remaining unchanged into the indefinite future? (Can I even rely on that?)
Wikidata's IDs are supposed to be stable.
- what is the policy on inclusivity? Do entities need to be 'notable' in
some sense to be accepted into Wikidata? (I'm imagining a research body wanting to offer very precise place or event data, or someone with the ambition to include in Wikidata details of any person who ever lived.)
https://www.wikidata.org/wiki/Wikidata:Notability
- is there a template for each entity type (e.g. person, place, event)
which guarantees that a query for certain properties will at least identify entities of the desired type? (My artist termlist query includes a test '$s ps:P31 wd:Q5' which picks out humans: I'm not clear how I would do the same for events or places.)
No that does not exist.
Cheers Lydia
Lydia,
Thank you for your responses. I suggest that it would add to the utility of Wikidata entities if their basic type ('person', 'place', 'event', etc.) was explicitly stated in the RDF.
Best wishes,
Richard
On 2016-06-29 4:56 PM, Lydia Pintscher wrote:
On Sun, Jun 19, 2016 at 1:27 PM, Richard Light richard@light.demon.co.uk wrote:
Hi,
My view is that, in order for there to be any point in cultural heritage bodies (museums, libraries, archives, historians) publishing their collections etc. as Linked Data, there needs to be a common Linked Data framework representing the historical space-time universe, which they can all quote. Current practice (such as the British Museum Linked Data offering) suggests that concepts such as people, places and events will otherwise be represented either by useless string values or by system-specific URLs which have no wider meaning.
As a result, I would like to explore the potential for Wikidata to act as this lingua franca for the cultural heritage community.
You'll see from my earlier messages to this list that I have been grappling with the SPARQL end-point. Initially I was confused by the interactive version of the Query Service [1], which differs in its response format from the similarly-URLed end-point and doesn't provide an RDF/XML response. I have now managed to set up Wikidata as a 'web termlist' service for artists, within the Modes software (see attached screenshot). (The data in the pop-up window is generated on the fly from the Wikidata RDF.)
At this point, I have the following questions:
- what level of stability is planned as regards Wikidata identifiers/URLs?
Can I treat the full URL (e.g.[3]) as persistent, or can I only rely on the core Wikidata identifier (e.g. [4]) remaining unchanged into the indefinite future? (Can I even rely on that?)
Wikidata's IDs are supposed to be stable.
- what is the policy on inclusivity? Do entities need to be 'notable' in
some sense to be accepted into Wikidata? (I'm imagining a research body wanting to offer very precise place or event data, or someone with the ambition to include in Wikidata details of any person who ever lived.)
https://www.wikidata.org/wiki/Wikidata:Notability
- is there a template for each entity type (e.g. person, place, event)
which guarantees that a query for certain properties will at least identify entities of the desired type? (My artist termlist query includes a test '$s ps:P31 wd:Q5' which picks out humans: I'm not clear how I would do the same for events or places.)
No that does not exist.
Cheers Lydia
Hi Richard,
I think that persons are probably most mature. All humans have instance of (P31) human (Q5) and for some subsets (like artists) a lot properties have a high coverage. Wikidata already connects to a lot of other sources on the web making it a central node for linked open data. For painters I keep track of this at https://www.wikidata.org/wiki/User:Multichill/Paintings_creator_no_authority... .
For events you probably need subclass of event, see https://tools.wmflabs.org/wikidata-todo/tree.html?lang=en&q=Q1656682&... For place geographic location is probably the best one, see https://tools.wmflabs.org/wikidata-todo/tree.html?lang=en&q=Q2221906&...
Also take a look at https://www.wikidata.org/wiki/Wikidata:WikiProject_Visual_arts/Item_structur... . We tried to link to external concepts here.
Maarten
Op 30-6-2016 om 12:07 schreef Richard Light:
Lydia,
Thank you for your responses. I suggest that it would add to the utility of Wikidata entities if their basic type ('person', 'place', 'event', etc.) was explicitly stated in the RDF.
Best wishes,
Richard
On 2016-06-29 4:56 PM, Lydia Pintscher wrote:
On Sun, Jun 19, 2016 at 1:27 PM, Richard Light richard@light.demon.co.uk wrote:
Hi,
My view is that, in order for there to be any point in cultural heritage bodies (museums, libraries, archives, historians) publishing their collections etc. as Linked Data, there needs to be a common Linked Data framework representing the historical space-time universe, which they can all quote. Current practice (such as the British Museum Linked Data offering) suggests that concepts such as people, places and events will otherwise be represented either by useless string values or by system-specific URLs which have no wider meaning.
As a result, I would like to explore the potential for Wikidata to act as this lingua franca for the cultural heritage community.
You'll see from my earlier messages to this list that I have been grappling with the SPARQL end-point. Initially I was confused by the interactive version of the Query Service [1], which differs in its response format from the similarly-URLed end-point and doesn't provide an RDF/XML response. I have now managed to set up Wikidata as a 'web termlist' service for artists, within the Modes software (see attached screenshot). (The data in the pop-up window is generated on the fly from the Wikidata RDF.)
At this point, I have the following questions:
- what level of stability is planned as regards Wikidata identifiers/URLs?
Can I treat the full URL (e.g.[3]) as persistent, or can I only rely on the core Wikidata identifier (e.g. [4]) remaining unchanged into the indefinite future? (Can I even rely on that?)
Wikidata's IDs are supposed to be stable.
- what is the policy on inclusivity? Do entities need to be 'notable' in
some sense to be accepted into Wikidata? (I'm imagining a research body wanting to offer very precise place or event data, or someone with the ambition to include in Wikidata details of any person who ever lived.)
https://www.wikidata.org/wiki/Wikidata:Notability
- is there a template for each entity type (e.g. person, place, event)
which guarantees that a query for certain properties will at least identify entities of the desired type? (My artist termlist query includes a test '$s ps:P31 wd:Q5' which picks out humans: I'm not clear how I would do the same for events or places.)
No that does not exist.
Cheers Lydia
-- *Richard Light*
Wikidata-tech mailing list Wikidata-tech@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech
On 2016-07-01 10:36 AM, Maarten Dammers wrote:
Hi Richard,
I think that persons are probably most mature. All humans have instance of (P31) human (Q5) and for some subsets (like artists) a lot properties have a high coverage. Wikidata already connects to a lot of other sources on the web making it a central node for linked open data. For painters I keep track of this at https://www.wikidata.org/wiki/User:Multichill/Paintings_creator_no_authority... .
Yes, artists look like a well-served part of the ecosystem.
For events you probably need subclass of event, see https://tools.wmflabs.org/wikidata-todo/tree.html?lang=en&q=Q1656682&... For place geographic location is probably the best one, see https://tools.wmflabs.org/wikidata-todo/tree.html?lang=en&q=Q2221906&...
I don't think a subclass is required: if anything the CIDOC-CRM definition of E5 Event is more abstract than the Wikidata definition of Q1656682. It's more a case of getting 'isa' (i.e. instance of) relationships into the data, so that a SPARQL query can know that it is only dealing with entities of the desired type. What is the Wikidata procedure/mechanism for doing something like that?
Also take a look at https://www.wikidata.org/wiki/Wikidata:WikiProject_Visual_arts/Item_structur... . We tried to link to external concepts here.
That's interesting. Has anyone made an attempt to map these properties to the CIDOC-CRM?
Thanks,
Richard
Maarten
Op 30-6-2016 om 12:07 schreef Richard Light:
Lydia,
Thank you for your responses. I suggest that it would add to the utility of Wikidata entities if their basic type ('person', 'place', 'event', etc.) was explicitly stated in the RDF.
Best wishes,
Richard
On 2016-06-29 4:56 PM, Lydia Pintscher wrote:
On Sun, Jun 19, 2016 at 1:27 PM, Richard Light richard@light.demon.co.uk wrote:
Hi,
My view is that, in order for there to be any point in cultural heritage bodies (museums, libraries, archives, historians) publishing their collections etc. as Linked Data, there needs to be a common Linked Data framework representing the historical space-time universe, which they can all quote. Current practice (such as the British Museum Linked Data offering) suggests that concepts such as people, places and events will otherwise be represented either by useless string values or by system-specific URLs which have no wider meaning.
As a result, I would like to explore the potential for Wikidata to act as this lingua franca for the cultural heritage community.
You'll see from my earlier messages to this list that I have been grappling with the SPARQL end-point. Initially I was confused by the interactive version of the Query Service [1], which differs in its response format from the similarly-URLed end-point and doesn't provide an RDF/XML response. I have now managed to set up Wikidata as a 'web termlist' service for artists, within the Modes software (see attached screenshot). (The data in the pop-up window is generated on the fly from the Wikidata RDF.)
At this point, I have the following questions:
- what level of stability is planned as regards Wikidata identifiers/URLs?
Can I treat the full URL (e.g.[3]) as persistent, or can I only rely on the core Wikidata identifier (e.g. [4]) remaining unchanged into the indefinite future? (Can I even rely on that?)
Wikidata's IDs are supposed to be stable.
- what is the policy on inclusivity? Do entities need to be 'notable' in
some sense to be accepted into Wikidata? (I'm imagining a research body wanting to offer very precise place or event data, or someone with the ambition to include in Wikidata details of any person who ever lived.)
https://www.wikidata.org/wiki/Wikidata:Notability
- is there a template for each entity type (e.g. person, place, event)
which guarantees that a query for certain properties will at least identify entities of the desired type? (My artist termlist query includes a test '$s ps:P31 wd:Q5' which picks out humans: I'm not clear how I would do the same for events or places.)
No that does not exist.
Cheers Lydia
-- *Richard Light*
Wikidata-tech mailing list Wikidata-tech@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech
Wikidata-tech mailing list Wikidata-tech@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech
Hi Richard,
I think that persons are probably most mature. All humans have instance of (P31) human (Q5) and for some subsets (like artists) a lot properties have a high coverage. Wikidata already connects to a lot of other sources on the web making it a central node for linked open data. For painters I keep track of this at https://www.wikidata.org/wiki/User:Multichill/Paintings_creator_no_authority... .
For events you probably need subclass of event, see https://tools.wmflabs.org/wikidata-todo/tree.html?lang=en&q=Q1656682&... For place geographic location is probably the best one, see https://tools.wmflabs.org/wikidata-todo/tree.html?lang=en&q=Q2221906&...
Also take a look at https://www.wikidata.org/wiki/Wikidata:WikiProject_Visual_arts/Item_structur... . We tried to link to external concepts here.
Maarten
Op 30-6-2016 om 12:07 schreef Richard Light:
Lydia,
Thank you for your responses. I suggest that it would add to the utility of Wikidata entities if their basic type ('person', 'place', 'event', etc.) was explicitly stated in the RDF.
Best wishes,
Richard
On 2016-06-29 4:56 PM, Lydia Pintscher wrote:
On Sun, Jun 19, 2016 at 1:27 PM, Richard Light richard@light.demon.co.uk wrote:
Hi,
My view is that, in order for there to be any point in cultural heritage bodies (museums, libraries, archives, historians) publishing their collections etc. as Linked Data, there needs to be a common Linked Data framework representing the historical space-time universe, which they can all quote. Current practice (such as the British Museum Linked Data offering) suggests that concepts such as people, places and events will otherwise be represented either by useless string values or by system-specific URLs which have no wider meaning.
As a result, I would like to explore the potential for Wikidata to act as this lingua franca for the cultural heritage community.
You'll see from my earlier messages to this list that I have been grappling with the SPARQL end-point. Initially I was confused by the interactive version of the Query Service [1], which differs in its response format from the similarly-URLed end-point and doesn't provide an RDF/XML response. I have now managed to set up Wikidata as a 'web termlist' service for artists, within the Modes software (see attached screenshot). (The data in the pop-up window is generated on the fly from the Wikidata RDF.)
At this point, I have the following questions:
- what level of stability is planned as regards Wikidata identifiers/URLs?
Can I treat the full URL (e.g.[3]) as persistent, or can I only rely on the core Wikidata identifier (e.g. [4]) remaining unchanged into the indefinite future? (Can I even rely on that?)
Wikidata's IDs are supposed to be stable.
- what is the policy on inclusivity? Do entities need to be 'notable' in
some sense to be accepted into Wikidata? (I'm imagining a research body wanting to offer very precise place or event data, or someone with the ambition to include in Wikidata details of any person who ever lived.)
https://www.wikidata.org/wiki/Wikidata:Notability
- is there a template for each entity type (e.g. person, place, event)
which guarantees that a query for certain properties will at least identify entities of the desired type? (My artist termlist query includes a test '$s ps:P31 wd:Q5' which picks out humans: I'm not clear how I would do the same for events or places.)
No that does not exist.
Cheers Lydia
-- *Richard Light*
Wikidata-tech mailing list Wikidata-tech@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-tech
wikidata-tech@lists.wikimedia.org
