On Mon, Mar 19, 2012 at 2:17 PM, Mark Rejhon <[email protected]> wrote:
> Version 0.2 of XEP-0301 (In-Band Real Time Text) has been released. >> [snip] > > URL: http://xmpp.org/extensions/xep-0301.html >> > [snip] > ChangeLog Summary: > [snip] > - Add "Multi-User Chat", "Simultaneous Logins", "Usage with Chat States", > and several other sections > >From testing my RealJabber software (demonstration XEP-0301 client), in terms of Chat States http://xmpp.org/extensions/xep-0301.html#usage_with_chat_states XEP-0085 (http://xmpp.org/extensions/xep-0085.html) specifies not to transmit <composing/> repeatedly. That said, many chat programs such as Google Talk (GMAIL) and Pidgin, Adium, etc, automatically assume <active/> for any <message/> transmitted without an XEP-0085 chat state. This is never properly clarified in XEP-0085 behaviour, about how to handle <messages/> not containing a chat state and what assumptions may safely be made. Therefore, XEP-0301 recommends transmitting a <composing/> with every <message/> that contains a <rtt/> but without <body/> to maintain backwards compatibility with existing chat software. Short summary: <message/> without <body/> and without any XEP-0085 chat state update, are still automatically assumed as <active/> (wrongly? or not?) by many existing software (i.e. GMAIL, Pidgin) which could be construed as wrong behaviour, but XEP-0085 does not seem to be directly clear about this specific situation. Since real-time text transmits <message/> without <body/> and with <rtt/>, and <rtt/> is real time text of the typing being composed, this is a wrong assumption to say <message/> without <body/> is to be assumed to be an <active/> state. A huge number of vendors seem to be doing this assumption, all the way from Pidgin thru GMAIL's Google Talk. It would be nice if XEP-0085 document can be updated to clarify in this area? Thanks, Mark Rejhon
