[Standards-JIG] Re: Roster Subscription Synchronisation JEPrevision 0.3

Richard Dobson richard at dobson-i.net
Tue Sep 14 15:33:44 UTC 2004


> The MSN protocol does support 'from' subscriptions. To understand how it 
> works
> you need a bit of knowledge of the MSN protocol, and here you go.
> (http://hypothetic.org/docs/msn for more information)

Yes I know all about the MSN protocol, I have made and implemented my own 
transport for MSN.

> If you don't understand that, you can take my word for it (as the MSN
> transport developer), that MSN does support 'from' subscriptions, and 
> every
> other type of subscription.

>From my previous observations when you deleted a contact from your msn 
contacts (in effect creating a "from" subscription state) the MSN client 
automatically adds that user to the block list (in effect turning it into a 
"none" subscription state), athough I have just double checked with the 
latest client and it seems to have changed so yes you are right it can have 
a "from" state.

> Also, the JEP I'm putting forward is not written simply to make my life as 
> a
> developer easier. I honestly do not think that any existing protocol is
> capable of fulfilling the requirements I've listed in the JEP.

But as I showed not all your requirements are valid as certainly quite a few 
of the use cases are already handled by standard subscription mechanisms.

> If somebody feels like writing up another proposal, then go ahead. I'm not
> disputing there are definitely other ways to fulfill those requirements, I
> just think that mine is the best.
> 1, because it will mostly work with existing clients (they'll receive 
> their
> list, and all the roster subscription updates)
> 2, it's not difficult to modify existing clients to be fully compliant
>
> If another proposal comes along that can fulfill the requirements better 
> than
> mine, I'll definitely consider using it. Currently, there are none.

I am planning on writing up a spec but will not be able to do so until 
around november, its best not to rush these things as you currently are, 
rushing things usually ends up with a less than ideal spec.

Richard





More information about the Standards mailing list