YAMP: ns-attr is a _mark_, not a _name_ for the namespace

At 06:56 PM 2000-06-01 -0400, John Cowan wrote:
>
>	xmlns:xsl="http://bailey.dscga.com/TR/xslt"
>
>As I have proved before, the fact that two URIs access identical
>entity bodies proves nothing either way, a fact quite independent
>of namespaces.  A URI identifies a resource, and a resource is
>not the same as an entity body.
>

Eureka!

The semantics of a namespace, In Accordance With the Namespaces in XML Rec,
is that all instances of the same namespace will associate some set of
QName's in the document with a particular [i.e. the same for all such
instances] ns-attr string by means of a namespace declaration in the
specified syntax. [I hope I quoted this back right.]

This does not make of the ns-attr a _name_ for the namespace; only a _mark_
which appears on all instances thereof.

Compare the distinction between Trade Name and Trade Mark.  The SHEFFIELD
or STERLING marks which appear on metal wares are not names.  They are not
nouns.  They are adjectives [ok, they are nouns, but they are elliptical
adjectival prepositional phrases, "from Sheffield" and "of Sterling
silver"], applied as marks.

A _name_ is a signification used, apart from the thing named, to refer to
the thing named.  A _mark_ is a signification applied to the thing marked
so you can recognize it or some characteristic of it.

The "namespace name" of the NS Rec. isn't "just a name;"  it isn't _even_ a
name.  It is an identifying _mark_.  This is a _mode d'emploi_ of
identification not contemplated within the range of "URLs are URIs and URNs
are URIs and for the Web they are one class."  The URI [schemes] all have
to follow the unified syntax because these identifiers appear jumbled
together in the Wild Wooly Web at arbitrary distances from the things to
which they apply or refer.  The ns-attr only applies or refers to the
corresponding collection of Qname's _right here in this XML instance_ and
no farther.  It can only appear in the sheltered environment of an
attribute value in a namespace declaration in XML; it doesn't need to be
hardened against the perils of appearing in sundry contexts ranging from
HTTP and MIME headers to coctail napkins.

No wonder all this fuss about URI generics sounds like overkill.

Al

Yet Another Modest Proposal

Received on Thursday, 1 June 2000 22:08:39 UTC