[Standards] geoloc: PEP=user, <iq>=resource

Stephen Pendleton pendleto at movsoftware.com
Fri Jun 1 15:25:43 UTC 2007

I didnt write the spec, but did a implementation of it. I agree that iq
method should return the location of the resource, and PEP the "user".

However, either I am going looney, or someone messed with XEP-0080.
Previously the example of the <message> response using the pubsub method

<message from='pubsub.shakespeare.lit'
         to='bassanio at merchantofvenice.lit'>
  <event xmlns='http://jabber.org/protocol/pubsub#event'>
    <items node='n48ad4fj78zn38st734'>
      <item id='i1s2d3f4g5h6bjeh936'>
        <geoloc xmlns='http://jabber.org/protocol/geoloc'>
  <addresses xmlns='http://jabber.org/protocol/address'>
    <address type='replyto' jid='portia at merchantofvenice.lit'/>

Now it seems to be:

<message from='portia at merchantofvenice.lit' 
         to='bassanio at merchantofvenice.lit/home'>
  <event xmlns='http://jabber.org/protocol/pubsub#event'>
    <items node='http://jabber.org/protocol/geoloc'>
      <item id='a1s2d3f4g5h6bjeh936'>
        <geoloc xmlns='http://jabber.org/protocol/geoloc' xml:lang='en'>

This seems to indicate that the <message> is sent with a from attribute of
the user, not the pubsub component - this seems to correspond to what PEP
specifies the behavior should be, but not pubsub. Also the <addrsses> tags
are gone. The more recent example seems to be wrong if you are using pubsub
(not PEP), because all the examples in the pubsub JEP indicate that the
message comes from the pubsub component not the user itself. Was this change
made to indicate that the geoloc node is for the location of the user in the
"from" attribute, or because there is an assumption that PEP is being used?
I was previously using the <addresses> information for this purpose.

As a larger issue I guess I don't see why there is a disconnect between
pubsub and PEP in terms of what the <message> from attribute is.

-----Original Message-----
From: standards-bounces at xmpp.org [mailto:standards-bounces at xmpp.org] On
Behalf Of Ian Paterson
Sent: Tuesday, May 29, 2007 8:46 PM
To: XMPP Extension Discussion List
Subject: Re: [Standards] geoloc: PEP=user, <iq>=resource

This new thread was supposed to be independent of the "overloaded 
resource identifiers" debate.

I'm not clear about what the geoloc spec intends. Could someone (e.g., 
one of the geoloc authors) please comment on the post below.


- Ian

Ian Paterson wrote:
> Different online resources might be in different locations. (For
> example if I leave my computer switched on at home while I'm at work.) 
> However the geoloc PEP protocol (see Section 4.1 of XEP-0080) does not 
> allow the publisher to specify which resource the location corresponds 
> to.
> IMHO XEP-0080 should therefore mention that when geoloc is published
> via PEP then it SHOULD refer to the location of the owner of the 
> account (i.e. the user), not the location of any particular resource 
> (i.e. the connection endpoint).
> The XEP might also want to specify that the  geoloc <iq/> response
> SHOULD refer to the resource - not necessarily to the user.
> What do people think?
> - Ian

More information about the Standards mailing list