Hi Arthur, There's no problem with supporting the syntax. My view though, is that there should be recommendations about the meaning of the various forms of a prefixed name. Beyond that, I do believe that there are forms, such as PN_PREFIX ':' or just plain ':', which don't have any useful meanings. Supporting syntax that has no useful meaning would seem to me to be just a cause for confusion.
Regards, Steve Pitschke Development Lead for IBM Rational Connector for SAP Solution Manager From: Arthur Ryman/Toronto/IBM@IBMCA To: Steven B Pitschke/Lexington/IBM@IBMUS Cc: [email protected] Date: 07/10/2012 11:08 AM Subject: Re: [oslc-core] Core Specification Version 2.0 Query Syntax: Prefixed Names Steve, Given that the SPARQL grammar rules do make sense and that they are part of the spec, what's the harm in supporting them? Regards, ___________________________________________________________________________ Arthur Ryman DE, Chief Architect, Reporting & Portfolio Strategy and Management IBM Software, Rational Toronto Lab | +1-905-413-3077 (office) | +1-416-939-5063 (mobile) |------------> | From: | |------------> >--------------------------------------------------------------------------------------------------------------------------------------------------| |Nick Crossley/Irvine/IBM@IBMUS | >--------------------------------------------------------------------------------------------------------------------------------------------------| |------------> | To: | |------------> >--------------------------------------------------------------------------------------------------------------------------------------------------| |Steven B Pitschke/Lexington/IBM@IBMUS | >--------------------------------------------------------------------------------------------------------------------------------------------------| |------------> | Cc: | |------------> >--------------------------------------------------------------------------------------------------------------------------------------------------| |[email protected], Arthur Ryman <[email protected]> | >--------------------------------------------------------------------------------------------------------------------------------------------------| |------------> | Date: | |------------> >--------------------------------------------------------------------------------------------------------------------------------------------------| |07/10/2012 10:09 AM | >--------------------------------------------------------------------------------------------------------------------------------------------------| |------------> | Subject: | |------------> >--------------------------------------------------------------------------------------------------------------------------------------------------| |Re: [oslc-core] Core Specification Version 2.0 Query Syntax: Prefixed Names | >--------------------------------------------------------------------------------------------------------------------------------------------------| I would object to disallowing predicates of the form ':comment' in OSLC queries. As Arthur mentioned, in our SPARQL and Turtle examples, we frequently use precisely this form with the prefix ':' defined as the URI for some local server - such as those generated for custom properties. So while the OSLC specs themselves do not need to use prefixed names of the form ':comment', a user of one of our OSLC providers might well want to do precisely this when writing a query for several custom properties. I would have no particular objection to removing support for PN_PREFIX ':' or just plain ':', though I question the value of the restriction. Nick. From: Steven B Pitschke/Lexington/IBM@IBMUS To: Arthur Ryman <[email protected]>, Cc: [email protected], [email protected] Date: 07/10/2012 06:35 AM Subject: Re: [oslc-core] Core Specification Version 2.0 Query Syntax: Prefixed Names Sent by: [email protected] Hi Arthur, Yes, I am exactly suggesting we restrict it to PrefixedName := PN_PREFIX ':' PN_LOCAL I am currently working on an implementation of the server side of query for the Lyo project. I will change my grammar to use the above, What is the procedure for changing the OSLC core spec? Steve Pitschke Development Lead for IBM Rational Connector for SAP Solution Manager From: Arthur Ryman <[email protected]> To: Steven B Pitschke/Lexington/IBM@IBMUS Cc: [email protected], [email protected] Date: 07/10/2012 07:57 AM Subject: Re: [oslc-core] Core Specification Version 2.0 Query Syntax: Prefixed Names Steven, All of your examples ( ":", "dcterms:" or ":comment") have clear semantics and are frequently used in SPARQL and Turtle. ":" means the URI associated with the prefix ":" "dcterms:" means just the URI associated with the prefix "dcterms:" Neither of the above is likely to appear in a query string. ":comment" is the concatenation of the URI associate with the prefix ":" and the local part "comment". This could be a valid predicate that might appear in a query string. In practice, no OSLC spec uses ":" as a prefix. Are you suggesting we restrict the grammar to only allow "reasonable" prefixed names? In that case, I believe the practice is to use prefixed names of the form: PrefixedName := PN_PREFIX ':' PN_LOCAL Regards, ___________________________________________________________________________ Arthur Ryman DE, Chief Architect, Reporting & Portfolio Strategy and Management IBM Software, Rational Toronto Lab | +1-905-413-3077 (office) | +1-416-939-5063 (mobile) From: Steven B Pitschke <[email protected]> To: [email protected] Date: 06/29/2012 01:15 PM Subject: [oslc-core] Core Specification Version 2.0 Query Syntax: Prefixed Names Sent by: [email protected] The core query specification http://open-services.net/bin/view/Main/OSLCCoreSpecQuery defers to the SPAQL specifction for Prefixed Names PrefixedName ::= /* see "SPARQL Query Language for RDF", http://www.w3.org/TR/rdf-sparql-query/#rPrefixedName */ However, the semantics and usefulness of some the allowed terminals are not clear to me: [68] PrefixedName := PNAME_LN | PNAME_NS [71] PNAME_NS := PN_PREFIX? ':' [72] PNAME_LN := PNAME_NS PN_LOCAL For example, what would the prefixed names of ":", "dcterms:" or ":comment" mean? I think it would be better to only support PrefixedName := ( PN_PREFIX ':' PN_LOCAL ) | PN_LOCAL which have fairly clear semantics. Steve Pitschke Senior Software Engineer and Architect Rational Software, IBM_______________________________________________ Oslc-Core mailing list [email protected] http://open-services.net/mailman/listinfo/oslc-core_open-services.net _______________________________________________ Oslc-Core mailing list [email protected] http://open-services.net/mailman/listinfo/oslc-core_open-services.net
