Currently blank nodes are not shown. The main reason is that they can’t be directly referenced in a SPARQL query.
It is probably not impossible but difficult. Here a few difficult cases and possible solutions:
- sh:property to a blank node sh:PropertyShape this could be indirectly expanded but what if we have more than on sh:property ?
- We don't really know wich node we mean to expand it indirectly with a SPARQL query
- As a workaround we can fetch the blank node with the sh:NodeShape. And expand it without a SPARQL query.
- This would work but what about chained blank nodes. sh:NodeShape - sh:property -> sh:NodeShape (bnode) - sh:path -> list .
- This can be workaround using CBD describe (stardog, qlever, ...) wich is a non standard way of describe.
- But what about incoming links from blank nodes? This is subject to investigate
Do we really need this? This is a crucial question.
-
Do we need to know in the graph view which list a node belongs to? I’ve already removed this case from the graph. The result without this was correct but strange.
-
In case of a sh:NodeShape it would be better do "show" the node shape in a certain way. But keep in mind. We often use sh:NodeShapes with sh:property sh:PropertyShape as bnodes. But Metaphactory is using sh:PropertyShapes as Nodes and sh:NodeShapes as bnodes.
But do we really need to see sh:Shapes in a graph explorer? It's only fun for RDF people but not for normal humans.
Beside that what are the usages of bnodes?
- lists - we want to see the list and not how a list is constructed in RDF (linked list)
- address node ?
- reification ?
- owl
There are many ways of showing valuable information about involved bnodes. But they are mostly related to a higher level concept. Maybe we should focus on them and not bnodes.
Currently blank nodes are not shown. The main reason is that they can’t be directly referenced in a SPARQL query.
It is probably not impossible but difficult. Here a few difficult cases and possible solutions:
Do we really need this? This is a crucial question.
Do we need to know in the graph view which list a node belongs to? I’ve already removed this case from the graph. The result without this was correct but strange.
In case of a sh:NodeShape it would be better do "show" the node shape in a certain way. But keep in mind. We often use sh:NodeShapes with sh:property sh:PropertyShape as bnodes. But Metaphactory is using sh:PropertyShapes as Nodes and sh:NodeShapes as bnodes.
But do we really need to see sh:Shapes in a graph explorer? It's only fun for RDF people but not for normal humans.
Beside that what are the usages of bnodes?
There are many ways of showing valuable information about involved bnodes. But they are mostly related to a higher level concept. Maybe we should focus on them and not bnodes.