AW: [Standards-JIG] avatars

Richard Dobson richard at dobson-i.net
Fri Jul 2 11:11:02 UTC 2004


> >They are not really in-band? They seem to be always sipub-ed.
> >
> >
> Yes. We dropped the idea while discussing.

Well they are no longer inband as in the data being inband, but I do think
inband is still appropriate as it has now moved to being a protocol for
specifying objects than can be retrieved using inband methods (i.e. via
jabber filetransfer methods, rather than via HTTP or something else).

> >But it shows that there are more applications for "getting data objects"
> >in different ways. I really would like harmonize all of them.
> >
> This was my intent, when I mentioned this proposal.
> Maybe we could unify it. Maybe it would suggest someting new to someone.

The one way in which it could be unified is using the iobj syntax in the
main pubsub node that I suggested in my previous email, then using the iobj
methods to retrieve the file using file transfer, IMO when alerting on the
change of an avatar the avatar data itself MUST NOT be sent, it MUST be pull
rather than push, especially since with avatars there is every possibility
that the user will already have it either as a standard avatar, or have
received it previously.

> >We should create a general protocol for getting data objects with the
> >options:
> >- in-band (plain and base64 encoding)
> >
> >
> We have JEP-0047.
> Why do you insist on creating another IBB protocol?

Yup there is no reason re-inventing the wheel, this is why we removed our
own inband mechanism in the first place.

> I really would not like to write another ibb_block module to my jabber
> server.

Nope certainly not.

Richard





More information about the Standards mailing list