[
Lists Home |
Date Index |
Thread Index
]
----- Original Message -----
From: "Jonathan Borden" <jborden@mediaone.net>
To: "Nicolas Lehuen" <nicolas.lehuen@ubicco.com>; "Leigh Dodds"
<ldodds@ingenta.com>; <xml-dev@lists.xml.org>; "Joe English"
<jenglish@flightlab.com>
Sent: Saturday, January 19, 2002 3:21 PM
Subject: What a namespace is.
> The primary feature of RDDL, and something that I think needs to more
> explicitly be stated in the next revision of the spec, is that RDDL allows
> the owner of the DNS entry which serves as the root of the URI to specify
> WHAT the namespace is, how the namespace will perform, generally what
> characteristics and properties it will have. In my talk at Extreme Markup
> 2001 I talked about the "Laws of Physics" for the namespace as something
> that the creator of the namespace may decide. Just as Newtonian mechanics,
> or Einstein's Relativity or String Theory etc. may be applied to our three
> or multidimensional SPACE, a nameSPACE is also a space, and the schemas it
> 'contains' serve as such Laws.
>
> YOUR namespace may not have any constraints, or it may be highly
> constrained. It may have a few resources or a billion. Your choice, you
get
> the benefits and accept the consequences.
Have a look at my fast attempt to explain how I see things :
http://nicolas.lehuen.com/Articles/Programming/Ideas/fog0000000033.html
I do think like you that a namespace is a space. I do agree that a namespace
can have laws, but I think they are very simple. The point where I may
totally disagree with you is that I think that the laws of the namespace are
NOT the laws of the documents you can build with names from the namespace.
The sheer existence of document mixing namespaces, beginning with RDDL
itself, proves that there are schemas that are above namespace laws. I think
namespace laws are very lightweight, basically allowing to verify whether a
name belongs to a namespace or not.
That is the reason why I think one should be very careful when associating a
namespace with meta-data like schemas or stylesheets. For laws that are
above namespace can't be wrote at the namespace level.
I regret, but I think that we will have to define a concept of document or
element type, to which various schemas are bound. Don't run away, a document
type name would be, like a namespace name, a URI. The laws of a document
type aren't defined intrinsincely, they are defined in different schemas, in
different languages.
A document type need not to be defined further more than just a rallying
point for schemas written in different languages. It has no inherent
structure ; the structure of XML documents belonging to a given document
type is defined by associated schemas, not the document type itself. A
document type is merely a handle, it's a structure space which floats above
name spaces.
In this context, I would be very happy to have a technology that allows me
to associate resources to a particular document type, like the example I
made in reply to Leigh :
<?xml version="1.0" encoding="iso-8859-1"?>
<?meta-data-catalog href="http://www.foobar.com/foobar.meta";?>
<foobar xmlns="http://www.foobar.com/namespace";>
<!-- real data here -->
</foobar>
instead of :
<?xml version="1.0" encoding="iso-8859-1"?>
<!DOCTYPE foobar PUBLIC "-//foo//bar//foobar"
"http://www.foobar.com/foobar.dtd";!>
<?xml-stylesheet href="foobar.css" type="text/css"?>
<foobar xmlns="http://www.foobar.com/namespace";
xmlns:xsi="http://www.w3.org/2001/XMLSchema";
xsi:schemaLocation="http://www.foobar.com/foobar.xsd";>
<!-- here begins the real data -->
</foobar>
> >
> > Yeah, let's get all human-readable resources out of the root RDDL
> document.
> > Why is it in the root RDDL document, anyway, whereas the DTDs, schemas,
> > stylesheets and other related resources aren't ? Just because it is
> supposed
> > to be fun to be able to type the namespace URL and read something about
> it.
> > Like many others, I don't think this is a sufficient justification.
>
> YOU can do this for your namespaces using RDDL. There is no requirement to
> have any human readable text in the root RDDL document. You _can_
reference
> resources which contain text in various languages using RDDL. You can
> reference other RDDL directories. RDDL allows this, on the other hand it
> does not mandate it, but it does not mandate much.
If RDDL only mandates that resolving a namespace URI does not gives an
error, then RDDL is not very interesting. If RDDL mandates that there is
some human-readable documentation available, then it is awkward but it can
work. If RDDL try to make us think that an RDDL document is a good place to
put reference to schemas, then RDDL is broken.
> 'Twisted conception" is an entirely subjective statement which is as of
yet
> unsupported by logical or scientific arguments that would allow one to
judge
> its validity. To be clear, I as well as a number of generally well
respected
> people in the XML community think that it can be _useful_ to have some
> association between namespaces and schemas (not necessarily a 1:1
> correlation however).
>
> I have spoke and written about this:
>
> [1] http://www.openhealth.org/talks/rddl.paper.xml
> [2] http://www.openhealth.org/talks/What%20is%20a%20namespace.doc
> [3] http://www.openhealth.org/talks/Extreme%20RDDL.ppt
> [4] http://www.openhealth.org/talks/XMLBioInformatics.ppt
>
> If you could respond to what is written in these papers, you will have a
> much better chance of demonstrating to me, why this is "twisted" rather
than
> "useful".
>
>
> My suspicion is that your concept of a namespace is entirely as a
syntactic
> way of indentifying XML elements.
That's exactly what I assume for now. It is why they were designed, and one
should not think of it another way for now.
Now, if you want to change this, why not ? Let's think about it. But don't
try to make us think that the laws that define documents schemas have to
reside in the name space. This won't work, and any RDDL document is a proof
it won't. Show me the algorithm. Using xml:base is not an option, but the
idea is there : why not drop the goofy idea that "a document type = a
namespace = a RDDL document = many schemas", use a special attribute like
XML Schema does and just write "a document type = a meta-data resources
directory = many schemas" ? You can even call this meta-data resources
directory RDDL, it will deserve its acronyms. For now it only deserves the
NDL acronym.
Imagine your reaction if the XML Schema group said 'documents will be linked
to their XML Schema document through the xml:base attribute'. Would you
agree ? Certainly not. So why are you proposing this for RDDL ? Drop the
namespace resolution system, use a specific attribute like XML Schema does,
and RDDL will begin to be usable, if only for RDDL document themselves.
> If so, there are many mechanisms of
> creating unique names that do not depend on URIs. In any case I claim it
is
> useful to consider namespace not simply as NAMES but as SPACES, either
> primordial spaces lacking structure, or as highly structured spaces. As
the
> legal owner of the namespace, you get to play the creator and devise your
> own Laws of Physics. Good Luck.
Think seriously about the name space structure. You won't find any schema in
there. You'd like to, and me too, but it's not that simple. Namespaces were
built to allow mixing elements with the same name but different meaning in
the same document. The resulting document can have a structure, it can have
a DTD, an XML Schema, RELAX NG schema, etc. But you won't find them as name
space laws.
Anyway, this is what I say BEFORE reading your papers. I'll read them now
and I am ready to change my mind if what you say is true. That's science.
> > Any claim of using RDDL to obtain schema information is a joke, because
it
> > assumes that a document has namespace which has a schema ! This is a
total
> > mess in concepts !
> >
> > I really don't understand why you smart people don't see that this point
> is
> > a problem. It's really simple to break it ! And once again, you just
> didn't
> > gave me the algorithm I was asking for. Try to think about it, you'll
see
> > that RDDL is really flawed.
>
> This is no joke. I am always happy to label my jokes. RDDL is not about
> schema validation algorithms (that's for schemas). Please read my articles
> and either accept what I say, convince me I am wrong, or we can agree to
> disagree. Note that RDDL does not require _you_ to associate a namespace
> with a schema.
If does not require me to do it. Fine. But by giving example that do it, you
just make people have false assumption about namespaces. I don't think the
debate about namespaces is closed ; but a sure thing is that RDDL is
promoting an interpretation of namespaces that I don't agree with, and I
hope you have read my previous postings sou you understand why.
Anyway. I think I can let go this RDDL thing and watch it explode. With a
bit of luck, it could generate an industry : "RDDL-compliant schema
validator !" "RDDL-compliant stylesheet engine !" "RDDL compliant J2EE
server with WebServices, ZeoSync(TM) 1-bit compression and all". People
would make money with it.
And after a while, we'll all burn those product on charges of over-hyping,
unusability, and short-sightedness, and I'll be able to push my technology
which will be slighty better, but not enough (because I wouldn't listen to
people pointing things that would not work), and it will eventually be
over-hyped, unusable and short-sighted too, etc.
Sigh.
Regards,
Nicolas
|
