[Standards] MSN does XMPP

Mark Rejhon markybox at gmail.com
Wed Sep 14 22:14:38 UTC 2011


Agreed...

I would like Microsoft to pass through XEP-0301 (In-Band Real-Time Text).
I know that someone from Microsoft's team contacted me inquiring about
this; an steps from Microsoft should be encouraged.

Standardized tests such as Acid made Microsoft Internet Explorer 9
much more compliant, so I think an "Acid" style test for XMPP might be
an interesting idea.  We need to raise XMPP compliance to a much
higher level.   Facebook's XMPP is also having some interop issues on
XMPP, it is rejecting certain protocols, too.

Someone would just slap together a standard XMPP "testing client" that
runs on two ends, and tests various XMPP protocols, ranging from
simple message deliveries, status changes, etc, to see if the system
goes through.   The two clients could bounce back results through the
clear XMPP messaging channel as a log (simple approach), or would also
talk to each other separately (directly or via a trusted Jabber
server) to "compare notes" (more complex approach).

Even a simplistic test client (1,000 lines of source code running on a
pre-existing Jabber library), could be slapped together by one of us.
It would only test just a few tests, and only brief undocumented
extension passing (can just be a small amount of random XML data in a
<message> payload to begin with) -- could be our "Acid Version 1" to
begin with.

Mark Rejhon

On Wed, Sep 14, 2011 at 1:43 PM, Nicolas Vérité
<nicolas.verite at gmail.com> wrote:
> Hi,
>
> Thanks for those of you who twitted it:
> http://www.liveside.net/2011/09/14/messenger-connect-is-now-live-connect-new-apis-for-skydrive-and-hotmail-calendar/
> XMPP Interface : You can integrate Messenger into your Web-based,
> desktop, or mobile instant messaging products by connecting to our
> XMPP service.
>
> It's Microsoft! Quite a sign... but...
>
> Now, what we may see is cheating on the protocol, interop risks.
>
> What do we have to prevent this? Nothing.
>
> Certification is far too complex and costly (for all).
>
> Simple tests (like ACID) may be a good way: with public testing and
> results, it lets the communities booh the cheaters.
>
> I think it's the most important task we have now.
>
> What is your opinion?
> --
> Nicolas Vérité (Nÿco) mailto:nicolas.verite at gmail.com
> Jabber ID : xmpp:nyco at jabber.fr
>



More information about the Standards mailing list