Re: [xml-dev] Are namespaces actually crypto-entities or crypto-links? (

[Arjun Ray <arayq2@gmail.com>]

 On Fri, 23 Jul 2021 13:21:59 +0200, Marcus Reichardt
<u123724@gmail.com> wrote:

| (<https://blog.jclark.com/2010/01/xml-namespaces.html>)

+1.
 
| My opinion is that namespaces were probably born out of the
| expectation that a wealth of new vocabularies would be designed for
| the Web, and hence a principled mechanism was thought needed for
| avoiding name collisions.

That was part of the post-facto "justifications".  The initial impetus
was the invention of "qualified names" in RDF-XML (a markup mishmash
from happy-go-taggy weekenders).  This required an imprimatur, since
it was part of the W3C Metadata Activity, which at that time Could Do
No Wrong and Could Not Be Gainsaid.  Qnames were a done deal,
essentially by fiat, and it was up to the XML Working Group to cook up
a suitable spec.

The business about name collisions was a remarkable episode in mass
delusion.  The means to avoid collisions were already known from the
ENR TC and the Hytime standard: they just weren't well-known, and once
everyone who mattered was pre-sold on Qnames as the greatest thing
since sliced bread, they remained not well-known for good.

[For completeness, here is an incomplete essay from long ago on how it
works: http://users.nyct.net/~aray/ns/ns.html.  And for how the ideas
fared on this mailing list, see
http://lists.xml.org/archives/xml-dev/200305/threads.html#00787. ] 

| it seems namespaces are on its way out, at least on the Web.

On Stack Overflow, I saw a comment remarking that "XML Namespaces are
cargo-cult programming at its finest."  Very succinctly put.

| So why not drop namespaces alltogether or at least have
| their definition not spill into parser layering with unwarranted
| complexity such as nesting and redefinitions etc eg. follow the
| approach of ISO-19757 (DSDL-9) and use eg.
| 
|     <?DSDL-9 bind-ns-to-prefix ns-iri="..." prefix="..."?>

I still think losing the atomic nature of basic tokens like names of
elements and attributes was a mistake.  But, as James wrote, what's
done is done.  The best we can do now is to ease the transition of
namespaces into obsolescence and eventual oblivion.