Andre-John Mas wrote: > > On 1-Nov-07, at 17:00 , Peter Saint-Andre wrote: > >> Ernest Nova wrote: >>> FWIW: >>> >>> http://googlesystem.blogspot.com/2007/10/google-to-connect-to-other-im-networks.html >>> >> >> Yes I saw that in my RSS feeds this morning. I don't have any inside >> information to share about it, but it's intriguing. However, it seems to >> me that they would prefer to do native XMPP with other services since >> that would give them real federation (not the client-proxy hack involved >> in traditional transports). > > Yup, from my experience few clients make the use of transports clear and > easy to understand. By ensuring everything is happening on the server side > it makes things easier for the average user. > > The only thing, is I am not sure I can imagine how they would deal with > MSN, > which allows for accounts with domain names which aren't in the msn.com or > hotmail.com realm. Maybe this could be a fallback method if everything else > fails, something like: > - if matches @yahoo.* try connecting using Yahoo > - if matches @icq.com try connecting using icq > - if matches @gmail.com try connecting using Jabber > - try other smart matches > - else attempt Jabber and on failure try MSN
Well, the closed IM silos don't use URIs or even DNS-based addressing standards. If they did, life would be much easier. For instance they could use im: URIs, and XMPP servers could resolve those per RFC 3861. For now we recommend that gateway addresses be of the form service.xmppserver.tld (e.g., msn.jabber.org). This is explained in Section 6 of XEP-0100: http://www.xmpp.org/extensions/xep-0100.html#addressing Peter -- Peter Saint-Andre https://stpeter.im/
smime.p7s
Description: S/MIME Cryptographic Signature
