Re: [xhtml2] Namespaces

On May 21, 2005, at 21:38, Anne van Kesteren wrote:

> Today I read:
>
>  <http://www.w3.org/2005/Talks/04-19-steven-XHTML2-XForms/>
>
> ...which confirmed some rumors about I heard about XHTML2 and XForms. 
> I guess XHTML2 will eventually adopt XForms 1.1 that has a new 
> namespace specified (using 2004 instead of 2002) and specifies some 
> kind of an "inheriting namespace".

As I see it:
1) Two vocabularies can, in theory, use the same local name. If these 
vocabularies are ever
    combined, there will be a collision.
    => Let's introduce namespaces.
2) The syntax for declaring namespaces is crufty and verbose and people 
don't tend to grok
    the syntactic abstraction of binding some insignificant prefixes 
locally to the
    significant URIs.
    => Let's introduce namespace inheritance which totally negates the 
usefulness of having
       namespaces in the first place, because now they can collide after 
all.
3) A vocabulary is overhauled, so lots of thing change (but a lot of 
things don't).
    => Let's introduce another namespace, which is like running Rot13 on 
all element names,
       so all app that have not been b0rked and have implemented 
namespaces correctly will
       have to change wherever element identity is compared.
4) Gratuitously changing the identity of all elements even if their 
semantics
    don't change is totally impractical.
    => Let's backport the useful stuff from the Rot13 (XHTML 2) 
namespace to the XHTML 1
       namespace and boycott the new Rot13 namespace.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Sunday, 22 May 2005 09:15:43 UTC