[Bug 1721] New: [FS] need a judgment to find the "best" type and quantifier

http://www.w3.org/Bugs/Public/show_bug.cgi?id=1721

           Summary: [FS] need a judgment to find the "best" type and
                    quantifier
           Product: XPath / XQuery / XSLT
           Version: Last Call drafts
          Platform: PC
        OS/Version: Windows 2000
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Formal Semantics
        AssignedTo: simeon@us.ibm.com
        ReportedBy: fred.zemke@oracle.com
         QAContact: public-qt-comments@w3.org


7.1.3 The fs:convert-operand function
The last three inferences under "Static type analysis" are inadequate
because Type1 is not deterministic.  What is wanted is that
Type1 is the minimal type such that Expr1 <: Type1.  For example,
suppose Expr1 is 1.3.  So Expr1 <: xs:decimal, but also 
Expr1 <: xs:decimal? and Expr1 <: xs:decimal+ and 
Expr1 <: xs:decimal*.  And many other types could go on the 
right hand side.  It wouldn't matter much except that 
quantifier(Type1) is required in the conclusion, and we don't know
whether the correct quantifier is 1, ?, + or *.  

The solution might be to create a judgment that Type1 is the 
best static type of Expr1.  Or the solution
might be to create a judgment that Quantifier1 is the best
quantifier of Expr1.  I notice that many other subsections of 7.1
could use such judgments.

This comment is related to two other comments:

1. #1612: We need to spell out the quantifiers in inferences
(is it "for all Type1" or "there exists Type1"?)

2. #1605: We need to define the inference engine (how does the 
engine know when it has computed enough inferences?  For 
example, "1.3 <: xs:decimal*" is correct, so can the inference
engine stop when it finds that one, and never find the tighter
inference that "1.3 <: xs:decimal"?

Received on Monday, 18 July 2005 21:31:18 UTC