[Standards] Re: XEP-0073: Question about service discovery

Rachel Blackman rcb at ceruleanstudios.com
Thu Feb 8 18:47:41 UTC 2007


> Speaking of presence packet sizes, this is what I get each time when
>  someone using Miranda changes their presence status:

I'd think Miranda is hardly the only client which does that.  I mean,  
at present, the vcard-temp avatar method is sort of the de-facto  
standard used by a number of clients.  (Including Google Talk and  
iChat, who I would argue are the local 800 pound gorillas.)  I'm an  
offender that way too, since I use the vcard-temp hash for avatars in  
Astra.

Even for clients that support JEP-0086, if they want their avatar  
able to be seen by older clients (or if they happen to log into a  
server that doesn't support PEP), they're going to have to include  
the vcard-temp hash anyway.  If they want to talk to REALLY older  
clients, they're going to have to include the x:avatar hash anyway.

I love talking idealized protocol as much as the next person, but I  
also have to live in the real world.  To use your 'gosh, that's  
spammy' example of vcard-temp and x:avatar...

If I do PEP-only avatars, and my users find that people on Google  
Talk, iChat and so on cannot see their avatars, those users are going  
to complain and log it as a bug against me.  /Me/, not the other  
clients.  And then I will implement vcard-temp hashes for Google Talk  
to see, and x:avatar hashes for Pandion users to see, and so on.

I don't dispute that PEP avatars are 'the right thing' in pie-in-the- 
sky perfect protocol Wonderland.  But unfortunately, I don't live  
there.  I really wish I did, but unfortunately, I don't think we've  
yet defined a XEP for the <white-rabbit/> stanza that can grab  
developers and drag them down the hole.

So at the risk of treading the same ground again (oh, look!  My  
footprints circa 2004... and 2005... and...), even if one supports  
PEP avatars, what incentive is there for client developers to /kill/  
support for vcard-temp and x:avatar at present, when it will  
inevitably and invariably lead to bug reports when some other client  
can't see the avatars (or when your own avatars stop working when  
you're on a given server that lacks PEP)?  Saying, "Well, that server  
doesn't support modern standards, not my bug," is just going to get  
the answer, "but when I log onto that server with iChat, people can  
see my avatar!"

I realize I'm probably on the verge of getting pelted with produce  
for mentioning this once more, but I'm going to point out -- again --  
that a yearly certification process might solve this.  In 2007, one  
could make PEP a 'recommended' feature for servers, and PEP avatars a  
'recommended' feature for clients.  Then in 2008, one could have it  
be a requirement; anyone who didn't implement it in 2007 doesn't get  
certified in 2008.  In 2009, you could even make a recommendation  
that clients /stop/ using vcard-temp hashes or they lose their  
certification.

But unless there is a push to /dump/ old protocol, even if PEP is the  
right thing and even if clients support PEP avatars, you're STILL  
going to see vcard-temp and x:avatar hashes in presence, just to be  
safe.  End-users (other than developers) rarely care about the  
protocol.  They don't care if their presence packets are 100 bytes or  
243 bytes... they just care that the Google Talk people can see his  
buddy Joe's avatar over on iChat, but not on your client, and they  
will log it as a bug.

-- 
Rachel Blackman <rcb at ceruleanstudios.com>
Trillian Messenger - http://www.trillianastra.com/





More information about the Standards mailing list