Re: [xml-dev] RE: James Clark: XML versus the Web

[James Clark <jjc@jclark.com>]

On Wed, Dec 1, 2010 at 4:11 PM, Michael Kay <mike@saxonica.com> wrote:

Namespaces are bad because they were done hastily without taking time to consider the consequences (or rather, without listening to the many people who were pointing out the consequences). They were a cheap hack, implemented at the wrong layer of the architecture, and without thinking carefully about the data model.

I agree it was done rather hastily, but at the time (as far as I remember) there weren't many objections.  It was not obvious how much pain they would cause.

I think the problem was that we started off with some major assumptions, which we never really questioned:

a) Documents using XML Namespaces had to be well-formed XML 1.0 (w/o namespaces)

b) Namespace names are arbitrary URIs

c) Most fundamentally, that we needed a "namespaces" standard

A major factor in assumptions (b) and (c) was the requirement to enable an XML syntax for RDF (which, ironically, seems less important these days).

James