Re: Namespace: what's the correct usage?

[Joel Rees <rees@server.mediafusion.co.jp>]

From a half-crazed newbie out in right field --

I was going to suggest that qualified nested names ought to mean something
entirely different from unqualified nested names. For example, in my mind,
<xsl:if> should instantiate a particular node as a particular xsl object,
where a plain <if> in the same place should instantiate an object of, uhm,
well, a piece of what is being instantiated, which would be whatever one
would otherwise expect at that node.

I, being used to nested contexts for some reason, would see nothing at all
unnatural in inheriting the expectation from the most recent ancestor node.
In fact, in my perception of perception, this is a necessary aspect of
perception itself. To back up my argument of perception, I was going to
offer the principle of inheritance in object-oriented languages, and the
corollary between a context-free grammar and a two-stack machine.

(I haven't kept up with this, but has anyone ever proven whether two stacks
is insufficient for un-restricted grammars?)

Now, Eddie, you point out to me that many people would deny my perception of
perception. Even if I am right and they themselves are managing their
perceptions through nested contexts, they may be unwilling to recognize it,
or even to understand what I am talking about. Then my recognizer branches
out and I consider multiple inheritance and my former manager who swore that
CoBOL could do anything C could because a procedural language is a
procedural language.

So I offer my two cents. But when I stretch forth my hand, behold, it is
empty.

Joel Rees

============================XML as Best Solution===
Joel Rees                          リース　ジョエル
Media Fusion Co.,Ltd.  株式会社メディアフュージョン
Amagasaki  TEL 81-6-6415-2560    FAX 81-6-6415-2556
    Tokyo　TEL 81-3-3516-2566  　FAX 81-3-3516-2567
                       http://www.mediafusion.co.jp
---------------------------------------------------
                                         Programmer
===================================================

Iwaku Eddie Robertsson <eddie@allette.com.au>


> > Sorry. I meant that an explanation along with this line is probably more
> > adequate for my "XML Schema: DOs and DON'Ts", as the reason of why one
> > should avoid unqualified local elements.
>
> Before this discussion comes to an end (if it ever will...) I think I'll
add my 2
> cents worth. I think we can agree that there is no "correct" way of using
> namespaces however both Simon and Kohsuke make a very good point that
using
> namespaces the way Martin does is confusing for people new to namespaces.
I've
> just spent 4 months writing a 2 day course on XML Schema (No it's not a
> comprehensive course that explains every detail). The course is intended
for
> people with knowledge in XML and XML Namespaces. I start off with a short
review
> of DTDs and Namespaces before going into XML Schema. Thus far I've only
had one
> session so this may not be very significant but what I expected to spend
time on
> was the more complex issues in XML Schema (complex type derivation,
abstract
> elements etc.) but I was surprised to see that the attendants understood
this very
> easily. Instead I had to spend a long time on the Namespace review and
especially
> default namespace declarations. As long as every element was prefixed it
was ok
> but default namespace declarations caused confusion. I had planned to
discuss the
> meaning of the elementFormDefault attribute but I ended up telling them to
always
> set elementFormDefault="qualified" in their schemas (at least in the
beginning)
> because there was no way I could explain it to them. I also suggested that
they
> always should use explicit namespace declaration because it is less
confusing.
>
> So, even though Martins use of namespaces is perfectly valid maybe we
should be
> careful with it's use in consideration of XML developers just beginning to
> understand namespaces.
>
> Cheers,
> /Eddie
>