Re: [xacml] Proposed semantics for operations involving INDETERMINATE

[Anne Anderson <Anne.Anderson@Sun.com>]

On 22 July, bill parducci writes: Re: [xacml] Proposed semantics for operations involving	INDETERMINATE
 > > B.  urn:oasis:names:tc:XACML:0.15i:operators:or
 > > 
 > >     1) Order of evaluation: not specified.  Operands may be
 > >        evaluated in any order.
 > >     2) When does evaluation terminate: when any one of the
 > >        following conditions holds:
 > >        a) One evaluated operand returns TRUE
 > >        b) All operands have been evaluated.
 > >           i)  And at least one operand returned INDETERMINATE
 > >           ii) And all operands returned FALSE
 > >     3) What result is returned:
 > >        2a)   TRUE
 > >        2bi)  INDETERMINATE
 > >        2bii) FALSE
 > 
 > 2bi seems to contradict 2a. in the case where i have 
 > 
 > <or>
 >   1=1 (T)
 >   1=2 (F)
 >   1=fred (I)
 > </or>
 > 
 > i can yield T (2a) or I (2bi). may i suggest that 2b read:
 > 
 > b) NO evaluated operand returns TRUE
 >    i)  And at least one operand returned INDETERMINATE
 >    ii) And all operands returned FALSE

Thanks.  That was my intent.

Anne
-- 
Anne H. Anderson             Email: Anne.Anderson@Sun.COM
Sun Microsystems Laboratories
1 Network Drive,UBUR02-311     Tel: 781/442-0928
Burlington, MA 01803-0902 USA  Fax: 781/442-1692