Re: Significant W3C Confusion over Namespace Meaning and Policy

On Tuesday, February 22, 2005, 6:14:42 PM, Claude wrote:

BCLL> So you don't think they are replicated to the local 
BCLL> machine along with the resources they specify even 
BCLL> if the machine is only occasionally connected?

Oh, I think they *are* replicated on the local machine. Its just that,
when you reconnect to the net, some of them might update. The whole
'dated/unchanging vs latest/changing' thing.

BCLL> Please pardon if I misunderstood your answer.

Sorry for not being clear.

BCLL> len


BCLL> From: Chris Lilley [mailto:chris@w3.org]
BCLL> Sent: Tuesday, February 22, 2005 10:45 AM

BCLL> On Tuesday, February 22, 2005, 5:33:03 PM, Claude wrote:

BCLL>> On the other hand, it should evolve down to a 
BCLL>> semi-stable set of stable DTDs or schemas that 
BCLL>> can be replicated.  Relying on an always connected 
BCLL>> system is dicey.  That is a recurring pattern of 
BCLL>> network design:  what to localize/replicate and 
BCLL>> what to centralize/connect to.  

BCLL>> Where do catalogs thrive?  It seems to be a best 
BCLL>> practice question worth thinking about.

BCLL> I think catalogs do best as a formalized and application-independent
BCLL> network cache.




-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead