On Tue, Oct 7, 2008 at 2:43 PM, Pedro Melo <[EMAIL PROTECTED]> wrote: >> There's also that the disco spec says that <identity /> should be >> consistent beyond what is being suggested here. > You mean section 6.3, Response Consistency?
I do. >> I think overloading identity at the deployment level is a bad idea, fwiw. > I like to have a name for each connection, and I don't think using the > resource is the sane way to do it. Perhaps, for my benefit (and possibly others who have trouble following these epic threads too), you could in quick bullet points summarise the arguments against static resource names. I know that the first is to avoid presence leaks - I'd debate that later, but for now I'd just like to understand the arguments against resources, as they do fit the bill for what we'd like them to do perfectly. > Identity name was a good candidate IMHO, > until Pavel mentioned the 115 cache-hit rate problem. Glad we've discounted that, then :) > Moving the name to a different spec is ok by me. Maybe even 198. If that's what ultimately needs to be done, I think a new spec could be done - I don't think it belongs in 198 really. /K
