Semweb4j/multipleModels
From semanticweb.org
Up | Previous lesson | Next lesson | Next step
Contents |
[edit] Lesson 3: Multiple Models
[edit] Questions & Answers
- Why use multiple models?
- you can split one big model into sub-models
- the submodels may allow reasoning while their union does not (because of contradictions)
- the submodels may contain different (or even disjoint) parts of the big model, which makes queries on the smaller submodels faster
- you can assign one model-per-user like del.icio.us
- you can create statements about models if you assign a context URI
- one model-per-statement allows statements about statements via Named Graphs
- you can use basic set arithmetics as union, intersection and complement
- you can split one big model into sub-models
- What are Named Graphs?
- A Named Graph is a model with an assigned context URI, so the whole model (or graph) is named.
- Because the "name" of the model is an URI, the whole model is also a resource
- The statements in Named Graphs are called quads (similar to triples) because they look like 'context, subject, predicate, object'. Named Graph powered triple stores are therefore sometimes called quad stores.
- What about Named graphs compared to reification?
- It seems that everybody has to decide for himself if he uses Named Graphs or reification to create statements about statements.
- What are SPARQL datasets?
- SPARQL datasets are almost like Named Graphs but there are a few differences.
- A SPARQL dataset can be empty.
- A SPARQL dataset must not but can have a context URI.
- What are quad stores?
- quad stores represent each statement as a quadrupel (subject, predicate, object, context).
- quad stores can't have empty sets of models.
- Which semantics follows the RDF2Go ModelSet?
- ModelSets are like SPARQL datasets.
- ModelSet contexts can only be URIs (no Blanknodes or Literals).
