W3C home > Mailing lists > Public > www-ql@w3.org > October to December 2003

Re: Initial value of statEnv.elemDecl and the static semantics of constructed elements

From: Mary Fernandez <mff@research.att.com>
Date: 21 Nov 2003 12:58:32 -0500
To: stijn.vansummeren@luc.ac.be
Cc: www-ql@w3.org
Message-Id: <1069437512.1509.27.camel@squeak.research.att.com>

Hi Stijn,

Good catch.  When statEnv.elemDecl(ElementName) is not 
defined, i.e., ElementName is not defined in any imported
schema, then the default type should be untypedAny.
There is a subtle difference between untypedAny and AnyType.
untypedAny is the type annotation used for well-formed
elements for which there is no additional type information,
whereas AnyType is actually  the top of the type lattice.

But I don't think that definition is in the current
document.  I will add it to our list of formal-semantics
bug fixes.  

Best,
Mary

On Fri, 2003-11-21 at 10:13, Stijn Vansummeren wrote:
> The static typing rule for constructed elements uses the "... in context ... 
> type lookup ..." judgement to retrieve the content type of an element name. 
> When the valiInitial value of statEnv.elemDecl and the static semantics of 
> constructed elemendation context is empty, this judgement depends on    
> 
>    statEnv.elemDecl(ElementName)
> 
> to retrieve the correct content type. 
> 
> My question is: what is the initial value of statEnv.elemDecl? The Formal 
> Semantics notes that this is the initial value of  the"in-scope element 
> definitions", which can be found at
>    http://www.w3.org/TR/xquery/#id-xq-context-components
> 
> There, the "in-scope element declarations" are defined to have "none" as an 
> initial value.  Does this mean that  statEnv.elemDecl(ElementName) is 
> undefined when no schema is imported (and hence, element construction fails 
> to type-check)? Or is it assumed that statEnv.elemDecl(ElementName) returns 
> AnyType on every ElementName (which gives every constructed element the type 
> AnyType)?
> 
> Best regards,
> --Stijn Vansummeren
> 
> Limburgs Universitair Centrum
> Universitaire Campus 
> Gebouw D
> 3590 Diepenbeek
> Belgium
-- 
Mary Fernandez <mff@research.att.com>
AT&T Labs - Research
Received on Friday, 21 November 2003 12:58:46 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:17:16 UTC