Hi, On Sat, Aug 27, 2011 at 4:14 PM, Ralph Meijer <[email protected]> wrote: > On Sat Aug 27 2011 07:24:14 AM CEST, Jehan Pagès <[email protected]> > wrote: > > >> Too bad the namespace does not follow the URN format, where we could >> simply update the version to change the hash definition. > > 'Simply updating the version' creates an entirely new namespace. There is no > magic here. Namespace URIs are mostly opaque identifiers. If we would need a > new 'version' I assume we would just switch to the URN-based scheme for > picking the new namespace. Practically you would still have to be able to > handle both if you want backwards compatibility, just like with Delayed > Delivery. >
Obviously. I am aware of these facts. But I think that with a versioning logics, protocol writers would take more care of keeping a close protocol, which makes evolution a lot painful for everyone (implementers first, then users who just know that "it does not work"). An "unrelated" new protocol on the other hand, the writers might feel they have no limitation to changes. What I would like to avoid is a situation with a completely new XEP having a completely different design, thus ending up with again a situation where we don't manage to see a clear decision of what has to be implemented, longing years. That's what I am afraid of and why I just say "ok if we can't find a backwards compatible algorithm, let's do a new XEP, but let's make the process as painless and transparent as possible please". Jehan
