Agreed. What matters (for versioning) is having a default processing
model that treats the elements with funny names differently. Right now,
attributes implicitly provide that segregation.
-----Original Message-----
From: Peter Hunsberger [mailto:]
Sent: Friday, July 15, 2005 8:24 AM
To: Andrew Layman
Cc: XML-dev
Subject: Re: [xml-dev] Mixed content in data-binding (Was: Re: [xml-dev]
Interesting pair of comments
On 7/15/05, Andrew Layman <> wrote:
<snip/>
>
> But there is also a versioning issue. Since adding such meta-data as
the
> last-modified date after the fact is a typical usage, attributes today
> allow such annotation with less disruption to existing readers. E.g.
>
> <email last-modified='2005-02-02'></email>
>
> But this only works to one level. One cannot add an attribute to that
> attribute.
>
> One could imagine a future version of XML that treated elements and
> attributes with more parallelism, and allowed attributes to have
> structure, perhaps with an element-like syntax, as in:
>
>
<email><@last-modified>2005-02-02</@last-modified></em
> ail>
>
> And
>
>
<email><@last-modified>2005-02-02<@says-who>Joe</@says
> -who></@last-modified></email>
>
Umm, not to be glib, but at that point you just have an element with a
funny name. Semantically, elements can already replace attributes, so
what's the point?
--
Peter Hunsberger
