[
Lists Home |
Date Index |
Thread Index
]
james anderson <james.anderson@setf.de> wrote:
| when the paper with the gory details appears, it would be nice if they
| were to define "external" and "local" with respect to "name" and
| "taxonomy" in terms of scope and extent with reference to process,
| document, document definition, and element.
I didn't want to write a formal treatise in a list message, so I used
terms which I *hoped* would be sufficiently clear in their apparent
meaning.
By 'taxonomy' I meant vocabulary or "namespace" or "tag set" - you know,
the stuff that goes between all those pointy brackets. A taxonomy is a
collection of 'name's in some coherent semantic scheme. (If people are
allowed to be vague with what a namespace "means", I'm not sure why
'taxonomy' can't simply be taken for what it *usually* means.) I chose
'taxonomy', because it has connotations of an organizing scheme, over
vocabulary, which suggests only a collection.
'Local' refered to the document in which the instance markup appears.
'External' refered to the provenance of some names. When you want to "use
the XLink namespace in an XHTML document", then the 'type' in 'xlink:type'
would be an external name, or a name in an external taxonomy, as distinct
from the local name (which could also be 'type', as it happens) given to
the *value* in the markup.
The scope is the document. One reasonable processing model is the way in
which an AF implementation could "extract" or "project" a virtual document
from the actual one, with all non-architectural, or non-this-namespace,
information removed.
| from (3) i infer that the extent of external names is indefinite and the
| scope of a local->external mapping is element, but it would be nice to
| see that defined.
| ? what is the extent of a local->external mapping and what is the extent
| of a local name,
I'm sorry, I don't understand what you mean by "extent". What would be
the extent, for instance, of markup like xlink:type="simple" in some
starttag?
| that is what is the relation to the infoset?
Per se, it's just a bunch of attributes. Theoretically, you could
*derive* infosets for implied virtual documents, factored by "namespace".
|
