RE: [Fwd: Re: CURIEs, xmlns and bandwidth]

Hi Pete,

> I think Mark's last message [1] is saying that - contrary to 
> my assumption below - the value space of CURIE is indeed a 
> set of {namespace name, local part} pairs, not a set of URIs.
> 
> But it might help the discussion if this is clarified?

Well, the idea of keeping the two parts separate is notional, really. I'm
only following existing practice to say that notionally we have an object of
type { URI, local-part } (not 'namespace name', as you said) that can be
mapped to a URI. The TAG note strongly suggests that this mapping is clearly
explained.

However, more importantly I think, I was trying to also say that I don't
believe that it makes any difference whether you have that intermediate
step. To me the problem is that the TAG note tries to have it both ways--it
says that QNames are *not* URIs, but if you have to use them as such, please
do so carefully.

The consequence of which is to leave too many gaps open, since it doesn't
give people an advised 'best practice' mechanism for abbreviating URIs. My
proposal makes no great claims to ingenuity (if only!), it simply tries to
formalise existing practice in a way that doesn't leave people hanging in
mid-air, and it does it by trying to move away from QNames (as in *proper*
QNames, rather than QNames-in-use).

Regards,

Mark
 

Mark Birbeck
CEO
x-port.net Ltd.

e: Mark.Birbeck@x-port.net
t: +44 (0) 20 7689 9232
w: http://www.formsPlayer.com/

Download our XForms processor from
http://www.formsPlayer.com/

Received on Friday, 4 November 2005 16:22:51 UTC