- From: Peter F. Patel-Schneider <pfpschneider@gmail.com>
- Date: Sat, 11 Mar 2017 12:02:54 -0800
- To: Irene Polikoff <irene@topquadrant.com>
- Cc: Holger Knublauch <holger@topquadrant.com>, "public-rdf-shapes@w3.org" <public-rdf-shapes@w3.org>
It's going to be hard. It's not possible to just remove the parts of the validation report that can vary because some of these parts have conditions on them. For example, removing type and subclass triples will prevent checking the SHACL instance requirements. There is also the problem that there are different RDF literals with the same value. Probably the biggest problem is that the number of values for sh:result can vary between SHACL Core implementations for the same validation. peter On 03/11/2017 09:30 AM, Irene Polikoff wrote: > Do you have any suggestions? > >> On Mar 11, 2017, at 7:56 AM, Peter F. Patel-Schneider <pfpschneider@gmail.com> wrote: >> >> The testing process is not adequate to check for compliance of validation >> reports. All that it takes to see this is to notice that there are multiple >> compliant non-isomorphic validation reports for when the data graph conforms >> to the shapes graph. Something more flexible than RDF graph isomorphism is >> needed as the central operation of determining whether the actual validation >> report is complaint with the requirements in the SHACL document. >> >> Peter F. Patel-Schneider >> Nuance Communications >> >> >> On 03/08/2017 11:52 PM, Holger Knublauch wrote: >>> I made some progress defining the test suite. >>> >>> http://w3c.github.io/data-shapes/data-shapes-test-suite/ >>> >>> Please take a look esp at the lower sections on how the comparison of expected >>> and actual results happens. >>> >>> This document was started by Jose who has since left the WG and since I have >>> no idea whether this still reflects his thinking I have inserted my name to >>> the list of editors. This does not mean that I am particularly keen on >>> "owning" this work or this document. It may not look too good if the primary >>> editor and implementer also defines the test cases. If anyone more neutral >>> wants to step up owning this task, I'd be more than happy to take a back seat. >>> Of course I will submit and edit tests though >>> >>> If this process looks reasonable, I suggest we start working on the actual >>> tests "soon", e.g. after reaching CR. >>> >>> Holger >>> >>> >> >
Received on Saturday, 11 March 2017 20:03:30 UTC