http://chatlogs.planetrdf.com/swig/2008-11-27.html#T10-53-15
My summary at this time:
Purl choice:
  - purl does not guarantee that the name will always refer to the  
same ontology. The owner of the purl could die, change his mind, sell  
his purl or whatever. There is no way for the purl organisation to  
know what would be correct and incorrect behavior
  + has the advantage of allowing one to change the url redirected to,  
so that those who do trust the domain owner can feel that taking over  
the
    purl will not be too difficult when an appropriate organisation is  
chosen
- http://purl.org/atom is taken, don't know by whome.
W3c url:
  + we have been proposed a w3c name space some time ago.
  - not sure what work remains to be done to put something there so  
that people can feel happy
    ( One thing Harry Halpin had tried to get us working on was a  
GRDDL for atom using XSLT 1. I have an XSLT 2.0 and Xquery transform )
  + this would make clear who the maintaining organization was
There are quite a few people on this mailing list btw, so I wonder  
what others make of this.
Henry
Case in point: http://purl.org/atom no longer points to anything, even  
though for a long time I think it did point to a whole bunch of atom  
related
stuff. So a purl is not much of a guarantee of anything.
Still I wonder what Sam would think about having that purl be the URL  
for the atom-owl namespace ?
> -Toby