Lists Home |
Date Index |
"Bullard, Claude L (Len)" wrote:
> From: james anderson [mailto:email@example.com]
> >"names of resources and names of locations of resources" - even
> >conflated - concern things which are independant of the one, single
> >property intrinsic to namespaces: the capacity to disambiguating names.
> >which, i would trust, remains part of the "core". establishing identity
> >with respect to
> >none of which has to do with either the names of resources or the names
> >of locations of resources.
> That is why namespaces shouldn't be in the core as long as
> they are tied to a system and are resolvable.
hmm... i saw nothing in either WD-xml-names11-20020403.html or
WD-xml-names11-req-20020403.htm or even in XML-SW-20020210.html which
stipulated that namespace names are "tied to a system and are resolvable".
> Identity isn't
> a property.
comparability is a quality of the namespace names. which quality is
essential to identifying names.
> Identification is a process.
which process has nothing to do with either tying the names of namespace
to a "system" or with resolving them. once the process runs its course,
it arrives at a determination, which is a quality or a condition.
> It's systemic.
in the case of universal names, in only the sense that a process which realizes
NSName e URI
Prefix e NCName
LocalPart e NCName
QName e ( (Prefix + NIL) x LocalPart )
UName e ( NSName x LocalPart )
NSContext = (Prefix + NIL) -> NSName
BindPrefix = ( (Prefix + NIL) x NSName x NSContext ) -> NSContext
InternName = ( (Prefix + NIL) x NSContext x LocalPart) -> UName
IsIdentical = ( UName x UName ) -> Boolean
is a system.
> We were beaten down with the "oh well, just because they
> can resolve doesn't mean we told them they could" arguments
> before. It leads to unreliable implementations and unpredictable
> results. Because the jury is out on almost everything beyond
> punctuation, there is no need to make a change now.
which is, as a system, neither less reliable nor less predictable than
one which resolves parameter entities.
namespaces are punctuation.
> Enough of
> the slippery slopes of minimalism.
is the goal, to avoid that with which namespaces are conflated, laudable
though it is, adequate reason to reject them in their entirety?