RE: [xacml] XACML standard identifiers

[Tim Moses <tim.moses@entrust.com>]

Colleagues - Why don't we debate and try to settle this on tomorrow's call?
All the best.  Tim.

-----Original Message-----
From: Seth Proctor [mailto:Seth.Proctor@Sun.COM] 
Sent: Wednesday, May 12, 2004 10:34 AM
To: Tim Moses
Cc: 'XACML'
Subject: Re: [xacml] XACML standard identifiers



On Sun, 2004-05-09 at 21:27, Tim Moses wrote:
> Colleagues - This came up in a side conversation at the face-to-face 
> ... I have defined all the identifiers as v2.0 identifiers.  I have 
> not re-used the v1.0 identifiers and their definitions.  The reason 
> for this is that the two specifications are not forward/backward 
> compatible.  And, all definitions that have not changed would have to 
> be omitted from normative sections of the v2.0 specification.  
> Naturally, if there is consensual preference, we can revert to v1.0 
> definitions.  What do people think?  All the best.  Tim.

Hi Tim. Sorry I didn't respond to this sooner, but I don't think I fully
understood what you were proposing until I saw the latest draft.

Personally, I think we should leave the 1.0 identifiers wherever possible,
especially if nothing has changed between 1.x and 2.0 for that identifier.
For one thing, it helps people understand what existed in 1.x and what was
added/changed in 2.0. For another, it makes it easier for implementations.
Also, what does this say about using 1.x identifiers in a 2.0 policy? Is
that legal?

That's my opinion, but I don't think it's a strong opinion, so if others
think we should rev all the identifiers, I'm probably ok with that. Just
want to make sure there's some discussion on this point :)


seth