On Tue, Jul 10, 2012 at 7:34 PM, Mark Rejhon <[email protected]> wrote: > Oh wow -- This is a massive surprise for me. > *precedent* in a *Final* standard -- that allows bypassing disco!
Before planning too much based on the precedent that CSN sets, it's worth noting that the precedent in question is that there was a protocol that predates caps and used another signalling method, and that this was retrofitted with caps support once it was available in preference to the legacy method, which was only left in place to deal with existing clients that deployed with the legacy method. That is: the precedent is that non-disco/caps discovery methods are deprecated. Also worth noting is that caps does not say that either method is appropriate. It mandates caps support, only allowing enough of the legacy support to allow interop with existing clients. I'm still aiming to do a review of the XEP soon - I'll try to suggest appropriate text for bits that I suspect need it when I do. /K
