You've probably at least seen:

http://florianjensen.com/2008/01/17/aol-adopting-xmpp-aka-jabber/

So I played with this a little yesterday .. very cool!  I certainly hope that AOL does indeed go through with this and embrace XMPP!  Based off my experience with OSCAR and XMPP, they -could- accomplish everything they're currently doing with OSCAR with XMPP.  Do they want to?  Is this simply a gateway to their AIM and ICQ services?  Who knows, but it's good to see!  I would imagine AOL has already seen how excited folk are about this, probably by their poor test server getting suddenly wailed on.  I can only hope they won't decide this is a bad idea.  The thing is, people like to use AIM and ICQ.  People like to use XMPP.  Why do they have to be at odds?  AOL's take on how chatting should work could bring wonderful improvements to XMPP overall.  I think Google's joining the fray did a lot of good for XMPP in general in terms of XEPs that came about because of it and such.  And people like me wouldn't be spending so much time building transports or multi-protocol clients if we didn't at some level like the services we were connecting to.  For example, I don't hate AIM.  I simply like a lot of protocols and don't want to run 8 different chat clients on my desktop.  I also think the OSCAR protocol is interesting, as are the other protocols, so I wanted to learn more about them.  Does that mean I'm out to make XMPP overtake AIM?  Hell no.  To be frank, it would be a happy day to me if the transports were no longer necessary, if people on AOL's servers could chat with people on my own server and on jabber.org's server and on lots of others without having to figure out the protocols.  Does this mean no one would try to dissect the protocols anymore?  Probably not.  Part of the drive of that is learning and understanding.  It doesn't have to be about trying to circumvent.

Imagine if AOL decided they really liked XMPP overall and ditched OSCAR though.  That would hose old clients.  But what if they released a new client that spoke XMPP while keeping their old servers going?  Furthermore, what if that client could point at other servers than AOLs.  Now AOL has a nice client or two out there for any XMPP clients to use.  That's kinda cool!  There's plenty of great XMPP servers out there at this point.  The client choices in general are a little lacking.  Spark, of which I'm in charge of nowadays, is pretty, it's cross platform, but it's a beast.  Java is not necessarily great in terms of small memory footprint.  Coccinella has a lot of cool features but it's Tcl/Tk and I've never liked the "feel" of Tcl/Tk.  Google Talk, there's not a version for my OS, so what good does it do me.  iChat...  nothing personal Apple but I've never liked it.  I can't put my finger on why really.  It's also hard to explain why I don't use Psi for everyday use.  I love it for development.  Nothing is better IMO.  But for some reason it didn't feel simple enough to be something I want to use as a regular client.  Adium X and Pidgin?  It's the primary client I use but it lacks a lot of XMPP functionality (though it's definitely improving on that front)  All of these are good clients, and have good followings, but there's still a lot of room for other 'entries' to come into the mix.  Nothing seems to "do it" for everyone yet.  Of course, who knows if that will ever happen.  But seriously AOL, I'd love to see you throw your hat into the mix!

Now what about MSN and Yahoo and others?  Yahoo, I could see them possibly embracing XMPP.  Hell they bought Zimbra and Zimbra's suite has XMPP capabilities nowadays, so they have some bought expertise there.  MSN on the other hand, they sell a product (LCS? OCS?  whatever it is) that may be a good drive -not- to embrace XMPP.  I don't know much about OCS or LCS though.

Overall though, this kind of reminds me of e-mail.  Everyone can have their own implementations of email services and such, that doesn't mean we can't all speak the same language!  AOL, I hope you take all this attention as a compliment!  I think we're all proud and pleased to see you express an interest in XMPP and maybe join the family!!!

So I just got done getting PyICQ up to a working condition, and I eliminated things that only AIM supports.  In the process, I discovered a lot of "overlap but not really" instances.  For example, AIM has the ability to change your password (well obviously), and so does ICQ, except it's not the same set of commands.  AIM has a weird directory search that doesn't work half the time.  ICQ also has a directory search, that doesn't work the same, and on top of that can be accessed in like 3 different ways per type.  (all dependent upon how you, the client, wants to pass the information to the server ... kinda weird to have that much choice)  Of course, using AIM's directory service protocol from ICQ works.  But accesses AIM's directory, not ICQ's.  It's a twisted array of bizarreness.  So if I want to do ICQ's search the "right way", I should redo it and not use AIM's.  (ICQ's is actually "inline", whereas AIM's you have to connect to a different server)

I ran into yet another instance of "I'm not announcing myself as a high enough version".  Again, it revolved around avatars.  Your typical bout of sitting there going "why isn't this working??? I'm doing the same thing as ICQ5!... oh crap".  Bah.  I want to do a lot of things to oscar.py, like make some handlers a lot more intuitive.  Right now, you have to actually know OSCAR a tad to use it.  I don't really like that.  For example, I'd like to be able to offer the ability for a person to add a user to their buddy list without having to know you are supposed to "start ssi mods, match up a group, add buddy to group, update buddy, update group, end ssi mods".  Should be a nice command "addBuddy".    Anyway, back to work.