Re: [wsbpel] Issue - 46 - Namespace for the document fragment representing a part

["Sanjiva Weerawarana" <sanjiva@watson.ibm.com>]

"David RR Webber - XML ebusiness" <Gnosis_@compuserve.com> writes:
> 
> I think you will find that your assertion at bottom
> on namespaces is slightly misleading.

I don't think so :-).

> We've been grappling with this exact same issue over in the
> CAM TC.  Here's what happens - if the XML content has a 
> default namespace declaration in it - then the XPath parser
> will demand prefixes, even if you don't use them!
> 
> This is an ugly bug and side-effect.

That's not a bug nor a side-effect. That's what namespaces
are .. if you use a default namespace that specifically means
that unqualified names are qualified by that namespace. 
Being prefixed or not has nothing to do with it ..

> The only recourse we have is to just warn people - do 
> not use namespace declarations - unless you absolutely 
> need them.

I agree default namespaces do confuse people.

> There are very few instances where you really need 
> namespaces in XML streams - paradoxically one is
> CAM templates - since they have embedded markup
> instructions to the CAM processor - and of course 
> these need to be discarded from the overall content
> processing.
> 
> Namespaces and their side-effects are something 
> that XML is going to have to live with for a longwhile
> downstream.
> 
> Fortunately tools like CAM remove the need to use
> namespaces in business transaction documents -
> therefore making their processing consistent and
> reliable.   Simple plain-vanilla XML is my preference
> everytime.

I disagree totally .. namespaces are a critical feature of
XML that makes XML scalable. However, for a reason that I'm
yet to understand, a lot of poeple confuse the presence or
absence of a namespace with that of a prefix. So as a result
namespaces have a "complex" bit associated with them .. but
I personally don't agree with that all. 

Sanjiva.