[
Lists Home |
Date Index |
Thread Index
]
Attributes have to be declared for the instance. PIs don't. As
to teaching, not me, fella. Too many big words. However, someone passed
this to me offlist:
http://lists.xml.org/archives/xml-dev/200003/msg00146.html
and it sums up what can be done with arch forms. I keep
thinking that using namespaces to disambiguate names and to
associate semantics is a bad move for the data itself, particularly
since it uses URLs (call 'em anything you like) because
the information is yoked to the system itself. That feels wrong.
At least with PIs, one is expressly saying, this is extra-curricular.
As to why I asked the question, it is more along the
line of, in the history of markup systems (and that is
a lot longer than the history of XML), we keep coming
back to this debate. Solutions are proposed but none ever
win. Why? Will a spec for solving this for XML make a
bit of difference? Maybe that is just a rhetorical
question but it seems that unless it is answered, we
will get to do this every year. Or just mail java
back and forth.
As to the complex names business, no duh, but I won't
kick them off the stoop for eating crackers. The guys
who did the AF work and Hytime work weren't fools even
if verbose. My guess is, there was peanut butter
on the crackers. Got milk?
Seriously, until we review all the historical approaches,
we keep going over the same old ground and just changing
names on the signposts.
Cowan's RNG post makes me think, what if the AF concept
were layered above the RNG? A clean separation
of validation and semantics seems to be what everyone
wants, but not one or the other exclusively. AFs are
expressed as PIs. Are they irrevocably yoked to DTDs
semantically? They appear to be looking over this example.
<?xml version="1.0"?>
<?IS10744 arch
name="somearch"
public-id="+//IDN me.com//NOTATION Some Architecture//EN"
dtd-public-id="+//IDN me.com//DTD Some Architecture//EN"
doc-elem-form="somedoc"
renamer-att="somedocnames"
options="option1 option2"
?>
<mydoc somearch="somedoc"/>
Steve, are they? Would archforms work with RNG?
len
-----Original Message-----
From: Nicolas Lehuen [mailto:nicolas.lehuen@ubicco.com]
Maybe the pile of attributes, though crappy, can give more meta-data about
an element than the architectural form ? Unless, of course, you have a
standard way to bind an architecture and its architectural form to other
types of meta-data than "just" DTDs. That would be great. That's why we want
to be able to bind document types to meta-data resource directories.
Beginner's question : what is the difference between an architecture and a
namespace ? Is it that architectures have a meaning (i.e. meta-DTDs and
possibility to write code that process the AF following these meta-DTDs) and
allow renaming, whereas namespaces are just sets of names ? Is all the
discussion about namespaces due to the fact that for many, namespaces should
have the same meaning as architectures (i.e. contain meta-schemas and allow
element and attribute renaming) ?
Disclaimer : I just read David Megginson's documentation of XAF [1], which
implements a subset of AFs for XML. I'll try to read [2] in the future,
especially for the meta-DTD part, but right now my brains just melt at the
idea of reading a document which calls its sections 'subclauses'. Apparently
those people like giving complex names to trivial concepts :P.
Yes, I know, Architectural Forms vs namespaces have already been debated on
this list, before namespaces where adopted, as early as in 1997 [3]. So now
that namespaces are there, what can we learn from AF ?
I begin to understand your question "Why is this debate still being held
every year?". It seems that everything has already been said and done, so
why should you care ? Well, I may have an answer that may give you some hope
: it's because some of us are here to learn, and you can teach us :) !
[1] http://www.megginson.com/XAF/
[2] http://www.ornl.gov/sgml/wg8/docs/n1920/html/clause-A.1.html
[3] http://lists.xml.org/archives/xml-dev/199705/msg00076.html
|
