More on Namespaces
david at megginson.com
Fri Aug 14 15:01:51 BST 1998
len bullard writes:
> David Megginson wrote:
> > The namespace URI is simply a unique prefix that can be shared by a
> > collection of names -- it is not guaranteed to point to anything at
> > all (some people wanted to *forbid* it from pointing to anything),
> Hmm. How is that better than having a Doctype root if subdocs
> were inlined? I can understand it not pointing to anything but
> only in the context of how FPIs work in Doctypes.
Personally, I'd like to separate the issue of global names from that
of document composition: I do not think that namespaces are a useful
substitute for subdocuments (inline or out-of-line) -- that issue
still needs to be dealt with. Namespaces simply allow an element or
attribute to have a name that can be compared across documents and
document types. The name is something that other facilities, such as
schemas or stylesheets, can act upon, but it has no significance in
In other words, an element named "http://www.megginson.com/ + para" in
one document and an element named "http://www.megginson.com/ + para"
in a second document have the same name. Make of it what you will.
> So what am I to assume if I see a dt: prefix that has a classID
> in the ns attribute, and another that *appears* to point at
> a schema of some sort?
I'm not certain that I understand the question -- could you give an
> If DCD isn't clearer than namespaces, than the HyTime drubbing
> goes down as one of the great crimes of critique.
> Precision aside, if the concepts and tools are harder to
> figure out than SGML and HyTime, this time next year I
> won't be the only one with a sore bottom. Wired and XavierMcLipps
> are watching. -)
The concepts underlying namespaces are simple enough (or could be,
with a little rewriting):
XML 1.0 elements and attributes have one-part names that are always
local to the document; namespaces allow the transformation of those
one-part local names into two-part global names, where the first
part is a URI and the second part is a simple name.
The WD contains much in the line of confusing and mostly-irrelevant
philosophical musings that (I hope) will eventually be deleted -- if
Len finds the spec confusing, I'm terrified of what will happen when
it hits the webmasters.
With the latest namespaces WD, however, XML can no longer fairly claim
to be simpler or more transparent than SGML -- the contextual
dependencies built into local scoping and defaulting are in the same
class as the contextual dependencies built into SHORTREF and OMITTAG
(both of which XML wisely discarded), though the algorithms for
resolving the namespaces are considerably simpler.
It is up to the reader to decide whether the fact of this complexity
vindicates HyTime or indicts XML. The temptation to obfuscate is hard
All the best,
David Megginson david at megginson.com
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;
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)
More information about the Xml-dev