[jira] [Created] (JENA-2089) RDFS for datasets

["Andy Seaborne (Jira)" <ji...@apache.org>]

Andy Seaborne created JENA-2089:
-----------------------------------

             Summary: RDFS for datasets
                 Key: JENA-2089
                 URL: https://issues.apache.org/jira/browse/JENA-2089
             Project: Apache Jena
          Issue Type: New Feature
    Affects Versions: Jena 4.0.0
            Reporter: Andy Seaborne
            Assignee: Andy Seaborne


This is not a replacement for any of the Jena inference and rules system.

"RDFS for datasets" is dataset and graph wrappers that take a vocabulary, build internal datastructures during setup, then apply the RDFS entailments to data, for both "match" ({{find}}) and "stream" (materialization) access to the data.

 * RDFS for datasets
 * Scale
 * Data interferences

{{ :x rdf:type ?type}} will returns the types of {{:x}}, deductions from {{rdfs:range}} or {{rdfs:domain}}, along with supertypes.

{{rdf:type}} will behave like {{rdf:type/rdfs:SubClassOf*}}.

Coverage: {{subClassOf}}, {{subPropertyOf}}, {{range}} and {{domain}}

It is fixed:
 * The RDFS vocabulary is not visible in the data and any vocabulary use in the data is not acted on.
 * The application can not subproperty the RDFS vocabulary (no subproperties of {{rdfs:subPropertyOf}}, {{rdfs:subClassOf}}, {{rdfs:range}}, {{rdfs:domain}} or {{rdf:type}}.
 * Vocabulary is static, not dynamically editable.
 * Inference in a dataset is "per graph", with the same vocabulary for all graphs in a dataset
 * The data can be updated
 * It is backwards chaining for scale.
 * There will be an equivalent Jena inference ruleset and test run both and compare the outcomes.

A "data+vocabulary" mode may be provided for graphs (i.e. data and vocabulary in the same graph), but not datasets which are split data and vocabulary only. Such a combined mode is not necessarily efficient - the focus of the work is split data and vocabulary.

The Dataset support will need JENA-2088.

In the future, incorporating directly into TDB1 or TDB2 evaluation, working with in {{NodeIds}}, should be possible.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)