[Standards-JIG] Re: privacy2 anti-SPIM proto-JEP

Tomasz Sterna tomasz.sterna at gmail.com
Wed Aug 31 18:26:33 UTC 2005

2005/8/31, Ian Paterson <ian.paterson at clientside.co.uk>:
> Instead of using the 'subscription' type, you add items to the list with
> type='group' for each group in your roster. (Or use type='jid' if these
> roster items are exceptions).

Wouldn't you agree, that isn't really setting rules via a subscription state?
I could really not use subscription and groups at all. Just add all
"jid"s to the rules.

> Obviously, the source of the difficulty here is that jabber:iq:privacy
> treats subscription='none' the same as 'not on my roster'.
> If you really think this is a serious problem in the wild (I'm not sure
> it is), then we could propose fifth subscription value (e.g. 'unknown')
> for jabber:iq:privacy in RFC 3921bis.

I don't really think it's a problem now.

But we have an item that describes really two different roster states
and it could be a problem in the future, so better raise it now.


More information about the Standards mailing list