Michael Rys wrote: >Contains expects a string typed value. The semantic rules of atomization >will employ fn:data()/dm:typed-value(). Even for fn:contains(). > >Your proposal is going to either remove useful functionality or add >additional complexity in the atomization process. Neither of which I am >in favor of. > > Or redefine xslt:value-of() and fn:contains(), acknowledging that they are text-oriented. For instance, fn:contains() could be defined using the following signatures: fn:contains($arg1 as item(), $arg2 as xs:string?) as xs:boolean fn:contains( $arg1 as item(), $arg2 as xs:string?, $collation as xs:string) as xs:boolean fn:contains() would return true if $arg2 occurs in fn:string($arg1). We probably would have to look carefully at our function signatures if we adopted my proposal, but I think this is a significantly cleaner and simpler proposal than the status quo. What useful functionality do you think this would remove? JonathanReceived on Thursday, 12 February 2004 15:07:10 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:14:05 GMT