chimezie: The Grammar is towards the bottom (the TOC links don't seem to work)
chimezie: The majority of the implementation against an abstracted RDF store happens here: http://cvs.4suite.org/viewcvs/4Suite/Ft/Rdf/Parsers/Versa/CoreFunctions.py?view=markup
chimezie: The grammar as BisonGen XML: http://cvs.4suite.org/viewcvs/4Suite/Ft/Rdf/Parsers/Versa/Versa.bgen?view=markup
chimezie: The majority of the implementation against an abstracted RDF store happens here: http://cvs.4suite.org/viewcvs/4Suite/Ft/Rdf/Parsers/Versa/CoreFunctions.py?view=markup
chimezie: The grammar as BisonGen XML: http://cvs.4suite.org/viewcvs/4Suite/Ft/Rdf/Parsers/Versa/Versa.bgen?view=markup
AndyS: Hopefully inline with the last call document, and current editors draft of the XML result format.
dajobe: As interest in a standardized RDF query language reaches a feverish pitch writes the author Chimezie Ogbuji
dajobe: I find this doc rather hard to follow
dajobe: forward traversal operator to start with
dajobe: versa contexts don't seem to be used the same as named graphs despite refering to them
dajobe: e.g Like their XPath counterparts, Versa contexts are associated with a set of variable bindings
chimezie: Versa contexts have a 'scope' associated with them. This is the 'name' of a graph to limit the query execution to (otherwise, every graph - named or otherwise - is part of the query execution).
dajobe: i've just worked out that A-B->C is meant to be an rdf triple arrow
dajobe: Traversal operators are almost direct, visual analogs of SPARQL's basic Triple Patterns [2.2] with Versa expressions at each part, hardly - they match only subject and predicates, not objects
dajobe: i.e just 2 parts of 3
dajobe: as the result seems to be a sequence of objects in the FTO
dajobe: although the description uses "context node" at this point which is puzzling
chimezie: the latter part (C in A-B->C form) does filter objects (via boolean expression). And objects of the matched statements are returned (as shown in the diagram: http://www.xml.com/2005/07/20/graphics/fwrd_traversal_operator_sm.gif)
chimezie: Subjects can be returned instead by using a different traversal operator ( A |- B -> C ). The only major difference (in function, not form) is the inability to return the properties matched in a traversal expression.
chimezie: My comparisons with SPARQL were meant primarily to ease the learning curve for those familiar mostly with SQL-form RDF query languages as well as to facilitate a birds-eye view on RDF querying in general
dajobe: I guess the style is more of a graph-query language than triple-query
dajobe: I got confused because the C in A-B->C is a different thing from the A and B
dajobe: versa lets you do subject,predicate queries returning objects with A-B->C or the reverse returning subjects
dajobe: although what | does I've not grokked yet
dajobe: I guess I get the idea of versa now. but the syntax twists my head, I can't remember what a particular fragment is querying or returning
chimezie: I was hoping the headache of a very different syntax (compared to the mainstream variants) would/could be aleviated by example usage (my motivation for writing Emeka).
dajobe: I find this doc rather hard to follow
dajobe: forward traversal operator to start with
dajobe: versa contexts don't seem to be used the same as named graphs despite refering to them
dajobe: e.g Like their XPath counterparts, Versa contexts are associated with a set of variable bindings
chimezie: Versa contexts have a 'scope' associated with them. This is the 'name' of a graph to limit the query execution to (otherwise, every graph - named or otherwise - is part of the query execution).
dajobe: i've just worked out that A-B->C is meant to be an rdf triple arrow
dajobe: Traversal operators are almost direct, visual analogs of SPARQL's basic Triple Patterns [2.2] with Versa expressions at each part, hardly - they match only subject and predicates, not objects
dajobe: i.e just 2 parts of 3
dajobe: as the result seems to be a sequence of objects in the FTO
dajobe: although the description uses "context node" at this point which is puzzling
chimezie: the latter part (C in A-B->C form) does filter objects (via boolean expression). And objects of the matched statements are returned (as shown in the diagram: http://www.xml.com/2005/07/20/graphics/fwrd_traversal_operator_sm.gif)
chimezie: Subjects can be returned instead by using a different traversal operator ( A |- B -> C ). The only major difference (in function, not form) is the inability to return the properties matched in a traversal expression.
chimezie: My comparisons with SPARQL were meant primarily to ease the learning curve for those familiar mostly with SQL-form RDF query languages as well as to facilitate a birds-eye view on RDF querying in general
dajobe: I guess the style is more of a graph-query language than triple-query
dajobe: I got confused because the C in A-B->C is a different thing from the A and B
dajobe: versa lets you do subject,predicate queries returning objects with A-B->C or the reverse returning subjects
dajobe: although what | does I've not grokked yet
dajobe: I guess I get the idea of versa now. but the syntax twists my head, I can't remember what a particular fragment is querying or returning
chimezie: I was hoping the headache of a very different syntax (compared to the mainstream variants) would/could be aleviated by example usage (my motivation for writing Emeka).