Re: DOCBOOK-APPS: Re: XSL: Docbook to foreign DTD

[Bob Stayton <bobs@caldera.com>]

On Sun, Jan 06, 2002 at 11:24:32AM -0500, Daniel Veillard wrote:
> On Sun, Jan 06, 2002 at 11:03:11AM -0500, Norman Walsh wrote:
> > Note, however, that the XSLT Rec requires only that processors output
> > sufficient namespace declarations to form a legal XML+Namespaces
> > document.
> 
> <pedantic>
>   Actually, it's far more stronger than that ! The rules about outputting
> namespaces declaration requires to always output a namespace when a node
> which had that declaration in scope is generated. I struggled with this
> until I understood that XML+Namespaces "correctness" is not sufficient
> to garantee that attributes containing CNames are gonna be preserved.
> 
>   Example: there is no way to know whether "ns:val" in the following
>     <foo xmlns:ns="http://example.org/bar" bar="ns:val"/>
> is actually to be consumed as a simple string, or as a CName value anchored
> in the http://example.org/bar namespace, assuming foo is to be copied as
> part of the output of an XSLT transformation, the rule is to actually serialize
> xmlns:ns="http://example.org/bar" to be sure that applications using 
> the namespace in attribute (or text) content do not break.
> 
>   Now I know why people were complaining loudly against early XLink syntax
> using CName in attributes ! The dependancy of the node toward the namespace
> is not expressed in any model (DOM, Infoset ...).
> </pedantic>

Could you clarify your example for me?  Does your "bar"
attribute have a declared attribute type of Nmtoken?
Otherwise the colon in the attribute value has no 
special signficance, right?


Bob Stayton                                 400 Encinal Street
Publications Architect                      Santa Cruz, CA  95060
Technical Publications                      voice: (831) 427-7796
Caldera International, Inc.                 fax:   (831) 429-1887
                                            email: bobs@caldera.com