[4suite] Namespace bug in 4DOM (Pretty)Print ?

[Uche Ogbuji]

Alexandre Fayolle wrote:

> In another mail I have sent this afternoon on xml-sig, I was mentioning
> accessing attributes created with setAttributeNS('',...) with
> getAttribute(...). I had overseen the strict division between the NS
> and the non-NS API.

Ah, I just replied to that message.  Never mind my reply, then.

> I strongly resent this (nothing on you, rather on the
> W3C in general). I feel some very awkward choices were made in the XML
> spec, and this causes very awkward backward compatibility issues that can
> be felt in the API.

You're not alone.  I like the spirit if XML 1.0 NS, but I also think
that some of the implementation is a severe pain.  However, I think some
of the choices were very hard ones made out of sheer necessity.  Just as
XML 1.0 had to struggle with its charter of SGML compatability, and some
"features" of XML are somewhat broken as a result, so too with the XML
1.0 group, which had to juggle issues of element scoping, established
XML parser practice, and the fundamental "meaning" of elements and
attributes inherited from SGML and HyTime through XML.  Then the DOM
group had to struggle with how to support XML 1.0 NS while maintaining
DOM Level 1 compatability.

Both working groups are frequently flamed for their choices on XML-DEV
and elsewhere but I'm not sure, considering everything, that I could
have made better choices myself.

As it is, if you adopt a particular set of practices, as we have based
on experience, you don't very often run into all the awkwardness.

Good luck.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji at fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python