- From: Michael Kay <mike@saxonica.com>
- Date: Fri, 15 Mar 2013 10:33:18 +0000
- To: Florent Georges <fgeorges@fgeorges.org>
- CC: Christian Grün <christian.gruen@gmail.com>, Jirka Kosek <jirka@kosek.cz>, EXPath <public-expath@w3.org>
On 15/03/2013 10:08, Florent Georges wrote: > On 14 March 2013 10:50, Michael Kay wrote: > >> Whichever of these is chosen, it should be clear that (like >> "numeric" in the F+O functions) it's not (necessarily) a real >> type that users can use in their own functions. > Would not that restrict the usability of such a type? So what are the options? (a) declare the argument as xs:anyAtomicType (b) accept only one of the binary types, and require the user to convert explicitly when necessary (c) have two variants of the function with different names (d) declare the function to accept a union type, which only works with a schema-aware 3.0 processor I can't say I like any of these much. Probably (a) is the least painful. Michael Kay Saxonica
Received on Friday, 15 March 2013 10:33:44 UTC