[Standards] Entity Capabilities Woes
melo at simplicidade.org
Wed Mar 14 22:14:28 UTC 2007
On Mar 14, 2007, at 4:16 PM, Tony Finch wrote:
> I'm somewhat baffled by the arguments against Matt's suggestion. It's
> basically the same design as used by many IETF protocols (SMTP, IMAP,
> POP3) which has worked quite well for over a decade. In particular,
> developers don't need to care about caps they don't support, and if
> are working on a new capability while it is under development
> (before it
> has a stable cap name) then they are already committed to following a
> changing specification.
I would say that it hasn't worked all that well. you need to register
to have a "short string" and that is a complex procedure. With caps,
I can develop something today and have it live on the net tomorrow
knowing that I won't mess up other clients.
If two IMAP server implementations decide to use XSEARCH for a new
kind of search, a client implementor cannot target that because they
might mean different things...
This kind of extensibility and low barrier for entry of new features
and protocols is extremely useful and powerful and I would hate to
let it go.
Jabber ID: melo at simplicidade.org
More information about the Standards