[standards-jig] LAST CALL: Service Discovery (JEP-0030)

Alexey Shchepin alexey at sevcom.net
Thu Dec 5 19:38:05 UTC 2002


Hello, Robert!

On Thu, 5 Dec 2002 09:26:51 +1100, you said:

 RN> I disagree - servername/admin (or servername/users and
 RN> servername/sessions, as my plan is for j2) is entirely appropriate - its a
 RN> different JID, so different results can be expected.
 >>  It works only because server not use this resources in another cases.  If
 >> I want to do somethng similar with client (or with service that connected
 >> as client) that already have resource, then I'll have troubles with using
 >> resource for this.

 RN> I'm a little confused by this. I'm talking about using the server JID, not
 RN> a user JID.

I'm talking about all JIDs, and solution that you use not suitable for all
JIDs.  Look below for example.  

 RN> Resource hierarchies are defined by the application.  The length
 RN> restriction is simply something that application designers need to take
 RN> into account. Resource partial matching is intended to assist the
 RN> application in creating such hierarchies, but it is certainly not the job
 RN> of the session manager to enforce any particular semantics on a resource
 RN> hierarchy.

E.g. clients can't send replies from any JID except user at server/resource.  So
if we want to make such hierarchy, then this will look strange that replies
comes not from JID they sended to...

 RN> I'm not entirely sure what this has to do with disco, either.

Look on "o/~ disco, disco MUC o/~" thread.  In short, the problem is: we have
conference room, e.g. jdev at c.j.o, and want to iq:disco it and get lists of
moderators, owners, banned users, etc.  But this look messy if we show all them
in one reply, better to return items with names like "Moderators", "Banned",
and get appropriate lists when discovery them.  How to do this with current
disco protocol?  We can distinguish these group by giving them another JIDs,
e.g. moderators#jdev at c.j.o, or jdev at c.j.r/[one or more not allowed for
participant chars]banned.  Do you like this way?  IMHO we send request to ROOM,
wich have JID jdev at c.j.r, and this is logically to have attribute
to='jdev at c.j.o' in IQ request.  But then we need to have standardized way of
grouping disco replies, and this can be easy solved by adding attribute in
<query/> and <item/> elements.

<newStrangeThought>
  Else lets remove xmlns attribute from IQ queries, and instead of it send them
  to e.g. jdev at c.j.r/jabber:iq:disco/moderators
</newStrangeThought>




More information about the Standards mailing list