- From: Peter F. Patel-Schneider <pfpschneider@gmail.com>
- Date: Tue, 22 Nov 2016 05:05:20 -0800
- To: "public-rdf-shapes@w3.org" <public-rdf-shapes@w3.org>
"RESOLUTION: Close ISSUE-196, deleting filter shape and instead, adding a boolean flag sh:disabled which (if true) means that a shape or constraint is ignored (i.e., any node is considered conforming/valid)." Ignoring a shape/constraint is different from treating it as if "any node is considered conforming/valid". Consider, for example se:s1 rdf:type sh:Shape ; sh:targetClass ex:Person ; sh:not se:s2 . se:s2 rdf:type sh:Shape ; sh:nodeType sh:IRI ; sh:disabled true . Ignoring se:s2 would probably result in se:s1 being uniformly valid. Treating se:s2 as uniformly valid would result in se:s1 being uniformly invalid. Which result is desired? Although it is true that filters can be replaced by OR/NOT, filters served a useful purpose. Removing filters makes it harder to write many useful SHACL constructs. I thus don't see why filters are being removed. Removing them does not appreciably simplify SHACL. Disabling a shape is not a replacement for filters. The resolution thus seems ill-formed. If filters are to be removed and a new feature is to be added, there should be two resolutions. Peter F. Patel-Schneider Nuance Communications
Received on Tuesday, 22 November 2016 13:05:54 UTC