[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: My report on experiments with unused namespaces

• From: "Costello, Roger L." <costello@mitre.org>
• To: "xml-dev@lists.xml.org" <xml-dev@lists.xml.org>
• Date: Wed, 15 Sep 2010 16:43:01 -0400

Michael Kay wrote:

> I actually came across one usage of XML in which the namespace 
> declaration in
> 
> <a xmlns:v="http://www.x.com/"/>
> 
> was considered significant. They were using it, rather as some people 
> use processing instructions or comments, as a device for adding 
> information to an XML document (in this case, a kind of version 
> attribute) without having to modify the schema.

What do you think about that? Is it a good idea? 

Until I saw that I was ready to pronounce all unused namespaces as evil, for the reason Simon sighted:

  I have heard of hacks which used namespace 
  declarations as kind of a backchannel for 
  propagating information 

But, in light of the above usage I must reconsider.

What do you think, is it best practice to delete all unused namespaces from an XML instance document? Or are there genuinely legitimate uses for unused namespaces? Is it a (bad) hack to employ unused namespaces as a device for injecting unchecked, out-of-band information into the XML document?

/Roger

• Follow-Ups:
  • Re: My report on experiments with unused namespaces
    • From: Michael Kay <mike@saxonica.com>
  • Re: My report on experiments with unused namespaces
    • From: David Carlisle <davidc@nag.co.uk>

• References:
  • My report on experiments with unused namespaces
    • From: "Costello, Roger L." <costello@mitre.org>
  • Re: My report on experiments with unused namespaces
    • From: Amelia A Lewis <amyzing@talsever.com>
  • Re: My report on experiments with unused namespaces
    • From: Michael Kay <mike@saxonica.com>