- From: Arthur Ryman <ryman@ca.ibm.com>
- Date: Thu, 19 Feb 2015 09:41:13 -0500
- To: Richard Cyganiak <richard@cyganiak.de>
- Cc: public-data-shapes-wg@w3.org
Richard, Yes, this is documented. The use cases come from Linked Data scenarios supported by OSLC. To describe the shape of an RDF graph it is important to specify if you expect URI nodes to be references to other graphs or to be subjects of triples in the same graph. OSLC has a property oslc:representation [1] which has values oslc:Inline, oslc:Reference, or oslc:Either. As an example, suppose a graph contains a table. Then the rows and cells of the table would be inline in the graph. However, the content of a cell might be a reference to another graph. We can model a set of graphs as a dataset in which the name of the graph is the URI whose RDF representation is the graph. That way we can avoid talking about HTTP and deferencing URIs. [1] http://www.w3.org/Submission/2014/SUBM-shapes-20140211/#representation _________________________________________________________ Arthur Ryman, PhD Distinguished Engineer | Master Inventor | Academy of Technology Chief Data Officer, Application Platform IBM Systems | Middleware 905.413.3077 (phone) | 416.939.5063 (cell) IBM InterConnect 2015 From: Richard Cyganiak <richard@cyganiak.de> To: Arthur Ryman/Toronto/IBM@IBMCA Cc: public-data-shapes-wg@w3.org Date: 02/19/2015 02:56 AM Subject: Use cases for RDF Datasets in shape validation? Hi Arthur, You said in yesterday's F2F discussions that you think of shapes as being evaluated against an RDF Dataset, not just a single RDF graph. Are there recorded user stories or use cases, or any other documentation, that motivate the necessity for this? My thinking so far was that validating single graphs is sufficient but I?d be happy to be convinced otherwise. Richard
Received on Thursday, 19 February 2015 14:41:48 UTC