- From: John Cowan <jcowan@reutershealth.com>
- Date: Mon, 29 Nov 2004 14:24:15 -0500
- To: Paul Grosso <pgrosso@arbortext.com>
- Cc: public-xml-core-wg@w3.org
Paul Grosso scripsit: > > One twiddle that isn't quite editorial that I'd like to see is > > banning attributes from the XLink namespace that aren't defined in the > > specification. Currently, if such appear they are undefined. > > > Can you outline some pros and cons to making this change? My defense would be a general one: the W3C owns the xlink namespace, and random users ought not to be allowed to create their own names with their own semantics in the namespace and at the same time claim conformance to XLink. Let them use their own namespace, or no namespace, for legitimate extensions. The addition of attributes to the XLink namespace should be made only by modifying or superseding the XLink Recommendation. > Why would you like to see this change? In what way is the > current behavior sub-optimal? What are the chances someone > is taking advantage of the current behavior and will be > negatively affected by such a change? I'd like to be able to not care about this, in the same way that we explicitly don't care about the way people may be currently using xml:id that would be invalidated by the xml:id recommendation when it is finalized. I certainly hope that no one is taking advantage of the ability to create xlink:this and xlink:that attributes. In that way, we can freely add attributes to XLink 2.0 without bothering about any existing conflicting uses. -- A witness cannot give evidence of his John Cowan age unless he can remember being born. jcowan@reutershealth.com --Judge Blagden http://www.ccil.org/~cowan
Received on Monday, 29 November 2004 19:24:48 UTC