Experimenting with Namespaces - DTDs?

[Steven R. Newcomb]

David Megginson (ak117 at freenet.carleton.ca) writes:

> Personally, I'd recommend architectural forms over namespaces if
> you're concerned with DTDs, since architectural forms have several
> major advantages:

David is right.

But I would go farther: XML Namespaces are a snare and a delusion.
With their use of colon syntax, they lull one into thinking that that
are about class inheritance.  They are not.  Instead, what the
namespace thing does is to collapse all the structure of the classes
of the inherited-from DTD into a salad of element types which is very
correctly termed a namespace rather than an architecture.  All that
RDF was looking for was a way to guarantee global uniqueness of
element type names, and if we ever try to get anything more than that
from namespaces, we are on very thin ice indeed.

If the inherited-from DTD is already a tag salad, in which all the
element types are a big OR group in the content model of the document
element, namespaces can work quite well.  If, however, an element type
has different meanings depending on its context (and most
architectures necessarily have this characteristic), then collapsing
such an architecture into a namespace can actively interfere with
information interchange.

I think RDF would benefit substantially, in terms of its
understandability, its implementability, and its flexibility, if it
were described in terms of inherited architectures.  In fact, I think
it cries out for an architectural perspective, in which the
knowability and significance of element context is preserved.  I
suspect that RDF's formal rigor would benefit, too, even though its
formal rigor is already formidable.  (I'm basically impressed by RDF;
it's the product of much excellent thinking, I think.  I just want
MORE!)

To be entirely fair and truthful, I must personally accept a share of
the blame for this namespace mess; I was present at the first Dublin
Core meeting, and, awed by the momentousness of the occasion, I
evidently failed to make the case for using architectures for
metadata.  My later contributions to the W3C XML discussions about
namespaces were evidently not persuasive, either.  In my own defense,
I would argue that this is entirely understandable; it's a subtle
issue; nobody has much experience with metadata architectures; what
experience there is is dominated by methodologies like MARC that rely
on lists of uniquely named fields; and, most of all, the need for even
a partial solution to the metadata problem is phenomenally intense.

Anyway, all is not lost.  This namespace thing is a mistake that will
necessarily be corrected, simply in order to support the needs of
civilization in an XML-dominated world.  The way toward a solution is
already paved by an ISO standard (ISO/IEC 10744:1997 Annex A.3) that
is being adjusted to accommodate the syntactic limitations of XML
(i.e., its lack of #NOTATION attributes).  It is implemented in the SP
parser and in other software systems, and it is already being used in
many industrial contexts.  It's the right sort of answer, it's not
going away, and its usage is accelerating rapidly; there was a
manyfold increase in the number of papers reporting its use at
SGML/XML 97.

And, anyway, the need for metadata interchange far outstrips RDF's
present scope.  I hope and believe that many powerful metadata
architectures -- including elegant ones that can't be squashed flat
and remain useful -- will be multiply inheritable.  That way, there
can be a marketplace of architectural ideas for metadata in which the
full power of context can be exploited.  I'd like to see RDF evolve
in that direction.

-Steve

--
Steven R. Newcomb, President, TechnoTeacher, Inc.
srn at techno.com  http://www.techno.com  ftp.techno.com

voice: +1 972 231 4098 (at ISOGEN: +1 214 953 0004 x137)
fax    +1 972 994 0087 (at ISOGEN: +1 214 953 3152)

3615 Tanner Lane
Richardson, Texas 75082-2618 USA

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)