Versioned Namespaces (was Re: [standards-jig] Avatars)
jajcus at bnet.pl
Thu May 8 07:55:28 UTC 2003
On Wed, May 07, 2003 at 11:57:35AM -0700, Justin Karneges wrote:
> My point is that you shouldn't be using experimental specs in the first place.
If experimental spec is not used it will never be well developed.
Experimental means something one can experiment with, probably to make
some good final protocol. And most effective are real-life experiments.
Asking not to use experimental specs is like asking to release only
fully-developed, fully-tested software - this would stop at least open
source software evelopement.
The poit is, that experimental specs should not be used any more, when the
final spec is available. At least unitl the final spec is implemented.
Namespace versioning seems a very nice solution to me. Every jabber
developer will notice when his software should be updated. And if he
doesn't do that the experimental implementation will not break anything
else. Final JEPs namespaces should not be versioned. And if a second
version of a protocol is being prepared its experimental versions will
be versioned again, but if the new final version is backward-compatible,
then it may use the same namespace as the first final version.
More information about the Standards