Re: Keywords, direct navigation, ...

[Eric Brunner-Williams in Portland Maine <brunner@nic-naa.net>]

Good morning everyone,

Something that troubles me in the model John proposes is scope.  Another,
and unfortunately related thing that troubles me is, as Bill noted in his
"reading ...?" note, is that this is not greenfields, and that symbolic
conflict with layers-as-markitecture is already in full swing.

The introduction of ontologies (policies associted to a scope-capable
mechanisn, if one is in the on-line ad racket), or categories (ditto),
or (venturing into the IDN mechanism minefield) character encoding schemes,
or (in venturing into the IDN policy minefield) partitions of namespaces,
is widely rejected "in the DNS", but (by a smaller, but significant set in
this BOF) accepted "above the DNS".

The same arguement holds that alternatives to the DNS that offer "direct
navigation" must, of necessity, have the property of ontological meaning.
The problem is, hosttables pre-existed the DNS, and were an alternative
to the DNS, as are (not were) resource-to-resource isomorphisms for several
alternatives to the DNS. ARP is an example.

I'm not convinced that an alternative mechanism (and here I _personally_
include dns-directory hybrid and directory models) of necessity has to have
a different scope than the dns (a set of names, a set of addrs, a mapping),
and has to import ontologies, artificial intelligence, facets, and lots of
other neat bits.

I don't want to repeat the IDN WG fiasco, all talk and no implementation.
For my $.02, this BOF must converge in the standard 2-tries to a WG that is
implementation-centered, not layerist. We keep ISO for that role.

Eric