[Standards] XEP-0160 and 'chat' messages

Matthew Wild mwild1 at gmail.com
Wed Jul 17 13:42:24 UTC 2013


Hi Ashley,

On 17 July 2013 12:58, Ashley Ward <ashley.ward at surevine.com> wrote:
> I am currently looking at implementing the rules around storing messages described in xep-0160 in openfire.
>
> The second rule states that messages of type 'chat' should be stored unless they only contain chat state notifications, which is fine. But I was wondering whether this could be generalised a bit more to perhaps suggest that any message of type 'chat' with an empty body should not be stored offline. xep-0334 could then be extended to allow a client to hint that a message should be stored offline
>
> The reason I suggest this is that there are newer xeps which also contain messages which you wouldn't want stored offline (e.g. xep-0301) and also unknown future ones, but which you also maybe wouldn't want to burden with a <no-storage/> hint.
>
> So essentially it makes the rule:
>
> 'chat' message stanzas without a <body/> should not be stored offline unless they contain a hint suggesting they should be.
>
> Can anyone think of any existing xeps which may be affected by this sort of change (e.g. ones which use a bodiless 'chat' message which would be expected to be stored offline?)

Not sure if you saw the recent discussion around this in relation to
XEP-0313 and XEP-0333 (MAM and chat markers)?
http://mail.jabber.org/pipermail/standards/2013-June/027701.html

Basically I think you are right, hence the hints XEP. What remains
uncertain is whether and when to default to storing or to not storing.

If there seems general consensus on what you propose (don't archive
things without a body by default, add a hint to make it archivable)
then I'll update both XEPs accordingly.

When considering this issue, it's important to remember that hints are
hints, and think about what happens if they are ignored (what
protocols will it affect if their body-less [disincarnate?] messages
are silently not stored?).

Regards,
Matthew



More information about the Standards mailing list