From: Vassilis Christophides ([email protected])
Date: 04/30/04
Hi Andreas Thanks for the mail. I would like to comment three of your criteria employed in RQL evaluation: 1) Recursion: Actually RQL supports recursion on the schema, e.g., select $X from Publication{$X} will return not only direct but also transitive subclasses of Publication. Another example is the basic query nca(Class1, Class2) returning the nearest common ancestor of Class1 and Class2 in a class subsumption hierarchy. On the other hand RQL do not support recursive data queries, e.g., select X,Y from {X}SubTopic*{Y} So RQL supports a restricted form of recursion according to your terminology. 2) Value Space: XML Schema datatypes are captured directly by the RQL semantics e.g., select X from {X}pages{Y} where Y = 8 and there is no need to cast strings to integers in queries ("8"^^<xsd:int>) So RQL supports Value Space queries for all XML base types 3) Collection and Containers: RQl not only provides the ability to query bugs and sequences, for instance, returned by nested queries e.g. select Y from (Person intersect Human){X}, {X}name{Y} but also to construct explicitly new container values e.g., seq(domain(pages), range(pages)[0] So I don't understand the justification of the restricted RQL support of containers in your report. Best Vassilis
This archive was generated by hypermail 2.1.4 : 04/30/04 EST