[Standards] Threaded chats

Robin Redeker elmex at x-paste.de
Sat Feb 10 00:07:42 UTC 2007


On Fri, Feb 09, 2007 at 12:55:02PM -0700, Peter Saint-Andre wrote:
> Greg Hudson wrote:
> 
> >Within a MUC, using <subject> to provide a human-readable thread topic
> >is likely to run afoul of the existing room subject behavior.  Although
> >we can mandate that new clients distinguish between messages with
> ><subject> and <body> and messages with only <subject>, it would probably
> >not interact well with old clients or MUC implementations.
> 
> Well, nothing in XEP-0045 as it stands today forbids or discourages the 
> sending of groupchat messages with <subject/> elements within the 
> conversation flow, and <subject/> is not reserved for use in MUC rooms 
> only to advertise the room subject. So IMHO if a client assumes that any 
> message with a <body/> and <subject/> received in the context of a MUC 
> room pertains to the room subject, then it is making a false assumption. 

XEP-0045 says:

   The subject is changed by sending a message of type "groupchat" to the
   <room at service>, containing no body but a <subject/> that specifies the
   new subject: ...

I agree so far :-)

But Overloading the meaning of tags like this gives me the creeps...
It's certainly not a clean solution, in my opinion at least.
Maybe the subject element that means the room subject and the subject
element that means the conversation element should have different names
or at least a different namespace.

When looking at it: the <subject> child of a <message> tag is clearly
the 'message subject' and not the "topic" of a room or anything else.
But I guess it's too late now to change such things :-/

It could've clearly been something like this:

   <iq
      from='wiccarocks at shakespeare.lit/laptop'
      to='darkcave at macbeth.shakespeare.lit'
      type='set'>
      <query xmlns='http://jabber.org/protocol/muc#topic>This is a great subject for a room!</query>
   </iq>

And then the muc does send the same to it's clients...

Aww... whatever, ignore my rants :-)


Robin



More information about the Standards mailing list