[Standards-JIG] Re: JEP-126 (Invisibility)

Trejkaz trejkaz at trypticon.org
Thu Apr 13 03:04:17 UTC 2006


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


On 13/04/2006, at 07:18 AM, Remko Troncon wrote:

> This sounds pretty advanced to me, and a protocol to get this seems  
> sounds very close to the full privacy lists management, so maybe  
> it's not too much to do anyway.
>
> Actually, the thing that makes privacy lists so awkward is the  
> chain of listing+retrieving+modding+submitting. I think that a way  
> to directly activate a submitted 'volatile' privacy list would  
> help. So, you do something like this:
> 	<iq type='set'>
> 		<query>
> 			<activate>
> 				<item type='group' value='myhiddenusers' action='deny' />
>      				<item action='allow' />
> 			</activate>
> 		</query>
> 	</iq>
> This will create a new privacy list on the server and immediately  
> make it the active list. When another list is activated (or the  
> client disconnects), the list is discarded. Such volatile lists  
> allow you to quickly set some kind of invisibility. However, this  
> leaves the question how permanently blocked users fit in this scheme.

That brings up an interesting thought... I was thinking this from the  
beginning, but as soon as you have a way to call one privacy list  
from another, it's starting to *really* look like iptables.

If we did have the ability to call one list from another somehow, and  
to append and remove items without having to transmit the entire  
list, then we would effectively be able to do this sort of stuff,  
right?  Then you just make a "blockedusers" list which always only  
has user blocking rules in it, and no matter what other privacy lists  
you happen to have in place, a client modifying the block list  
wouldn't interfere with your main rules.

TX

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (Darwin)

iD8DBQFEPb+2uMe8iwN+6nMRAi26AJ91a+Gd3a7gyzynyhV/ic9G1htlJgCcCnD2
ty9qO7Md5jLimY0SD4pQsTo=
=X47P
-----END PGP SIGNATURE-----



More information about the Standards mailing list