xml-dev - Are Namespaces Broken?

[Wayne Steele <xmlmaster@hotmail.com>]

>From: Rick JELLIFFE <ricko@geotempo.com>
>
>I have a modest suggestion: no-one should ever say anything is broken
>again unless and until they also have a working proposal of how to fix
>it. If we all kept to this discipline, XML-DEV would be more stimulating
>for everyone, and lead into a more beautiful future.
>

I have proposals on how to fix namespaces, but proper implementation 
requires a time machine. I don't think the Namespaces rec is hopelessly 
"broken" - If I did, I probably wouldn't waste my time discussing it.

I think probably the best way to move forward is with standardization and 
implementation of Schema, DTDs become secondary and in practice deprecated, 
and other XML-family specs (infoset, DOM, XSL) shift to harmonize with the 
new world of XML Schema, but not so much as to become irrelevant to XML 
without schema.

Nothing controversial about that.

I still think it's important to discuss the major shortcomings of 
Namespaces. We need to hash out "where things went wrong" so we can avoid 
the giant potholes that rec left in the road for us. So we can be alert to 
confusion (wrt interpreting the Namespaces Rec) from Vendors, XML Newbies, 
or W3C Working Groups.

What are these "back-in-time" proposals I have about Namespaces? Yes, I 
think it's too late now. But still...

1. The Namespaces Rec says it is not a goal for the URI to have to point to 
a schema (or something), and specifically mentions URNs as a good idea. Then 
this same rec proceeds to define
   http://www.w3.org/XML/1998/namespace
as the namespace which the "xml" prefix binds to!

Why, for the love of God, Why?

Couldn't it have been URI://w3.org/XML/namespaces ? What does the Hypertext 
Transport Protocol have to do with this?

2. The interaction of Namespaces and DTD Validation was not only not 
harmonized, it was never discussed in the rec - almost as if it had never 
occured to the authors that there was an issue. The result is that some 
Parsers STILL validate differently than others.

3. The DOM people knew damn-good-and-well about the namespaces rec, but they 
ignored it like a pariah. While somewhat understandable, this did great harm 
to both the acceptance of Namespaces and the usefulness of the DOM.

I hope I haven't offended anyone with these observations.
The people who "did this to us" are all smart, honest people.
I just would like them to take a little nibble of crow for how it turned out 
- to say "In hindsight, we should have done it differently."


-Wayne Steele



________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com