Re: [ubl] PRD3 suggestion: sample instances with unconventional namespace prefixes

["G. Ken Holman" <gkholman@CraneSoftwrights.com>]

At 2011-04-02 20:16 +0200, Roberto Cisternino wrote:
>the use of suggested prefixes (cbc, cac, ...), even if not normative,
>helps a lot for many reasons:

I *totally* agree, Roberto.

>1) If we are going to use SOAP envelopes and WS* there are several
>namespaces that are in place along with UBL namespaces, so it is very
>useful to use those prefixes suggested on each specification (even if not
>normative)

Useful, yes, normative, no ... I agree.

>2) by relying on automatic prefixes (n1, n2, nx, ...) generated by tools
>it is really risky.

Oh?  It shouldn't be risky in any manner whatsoever.  But I'm not 
sure what you mean by "relying on automatic prefixes".

No-one should rely on *any* prefixes ... automatically generated or 
authored.  *That* is the point of my posting the sample documents 
with unconventional namespace prefixes.  They aren't "wrong" and the 
instances are not invalid ... simply unconventional.

Conforming XML-aware software handles *any* correct use of 
namespaces, conventional or unconventional.  If a vendor cannot 
handle the sample instances that I've posted and treat them as fully 
conforming UBL documents, then the vendor's software is not 
XML-namespace-conformant.

>In PEPPOL we experienced real troubles with JAXB which
>is seriously bugged with the xs:Any element handling and it is very easy
>to have duplicated namespace prefixes when you use it.

I'm basing my comments solely on the XML and XML Namespaces specifications.

Any UBL implementation has to be prepared to accept *any* correct use 
of XML Namespaces, otherwise it will be left behind in 
interoperability within the user's community.

I cannot comment on whether or not JAXB is XML-namespaces-conformant, 
as I've never used it.  I would be very surprised if it didn't 
conform, but perhaps the tool creators misunderstood the use of namespaces.

>3) UBL namespace prefixes are providing a more readable document.

Totally.  And when humans read a UBL document, using the documentary 
conventions we've used in our materials will make it easier for that 
human to understand what they see.  It will be more "readable" as you suggest.

But a computer doesn't care what the namespace prefixes are, which is 
exactly my point.  "Readability" isn't an issue for computers, 
accuracy is.  Our UBL users' software must not care what the 
namespace prefixes are in the document.  Hence, those sample 
instances I'm suggesting we add to the distribution will equip users 
who hear unfounded claims of conformance from vendors.

Roberto, throughout the project I've been recommending for the 
readability by humans of the specifications that the UBL committee 
adopt namespace prefix conventions for clarity ... here are just a 
few citations from way back:

   http://lists.oasis-open.org/archives/ubl-lcsc/200306/msg00077.html
   http://lists.oasis-open.org/archives/ubl/200410/msg00002.html
   http://wiki.oasis-open.org/ubl/Example_Calculation_Models#ns

Absolutely we continue our committee's convention of documenting UBL 
for human eyeballs using a convention for namespace prefixes.  These 
will keep our documents readable by human readers.

But absolutely we must not prevent a perfectly conforming UBL 
document from being machine processed, regardless of what namespace 
prefixes happen to be used.

Thanks for your comments.

. . . . . . . . . . . Ken

--
Contact us for world-wide XML consulting & instructor-led training
Crane Softwrights Ltd.          http://www.CraneSoftwrights.com/o/
G. Ken Holman                 mailto:gkholman@CraneSoftwrights.com
Legal business disclaimers:  http://www.CraneSoftwrights.com/legal