July 02, 2006

Intelligence at which edge?

Reviewing Ken Camp’s article on Microsoft’s unified communications strategy, I came across this remark attributed to Bill Gates:

The problem is that our communications identities and experiences are linked too closely to our location, our devices and the mode of contact we are using. Your work number is tied to the phone on your desk. Your cell phone number calls the device you carry in your pocket. You may have separate identities for email and instant messaging, plus a number you call for audio conferencing and a code you must input.

This is far too complicated. Unified communications will reduce complexity by putting people at the center of the communications experience. Our goal is to integrate all of the ways we contact each other in a single environment, using a single identity that spans phones, PCs and other devices.

No! Please, no! People want to have separate identities, but attached to their different contexts; devices and geographies just happen to map partially to those contexts. Indeed, the system we’ve had for decades (call me at work for work issues, at home for personal ones etc.) has worked reasonably well. We’re trying to optimise a system that is only moderately disfunctional, not one that’s totally broken. That’s why iotum’s stuff works! They’re not trying to make everyone change their behavior.

But there’s a way, way deeper problem. Services like voicemail need to be totally re-thought for this world. One of the insights of the end-to-end principle that animates the design of the Internet is that only the edges have the context with which to add value and make decisions about how the application should behave. But when the communication is between two peers, who really decides what should go to voicemail? Which of the two edges does what?

In my ideal future world, the timeout on a call ringing would be set by the minimum overlap of the preferences of the caller and callee; since the callee is the one being interrupted without any choice, theirs take precedence.

When the timeout occurs, the callee’s telephony system redirects the caller to voicemail by handing back to the caller an appropriate URI and access tokens. Note that it’s the caller’s decision to leave a voicemail; after all, you can lead a voice to voicemail, but you can’t make him speak. Now the caller’s user experience may come to vary here; the call might have been tagged beforehand as a “get Bob RIGHT NOW” by pressing the green call button 3 times, in which case voicemail will be ignored and it’ll try his mobile instead without user intervention. Or you might get a locally-generated IVR with “press 1 to leave Bob a voicemail”. Indeed, that might be in your own language — how frustrating is it to call a colleague abroad and have to talk to an IVR in a foreign language!

That’s not unified communications, however, which is a supplier-centric solution for a problem the users simply don’t have. It’s only by the edges co-operating in new ways that we can truly add value. Intermediating the system by making it smart in the middle is ruining its semantics: a late-night call to Bob’s office phone with the intent of leaving a voicemail might accidentally set off his home phone ringing at 3am. Microsoft has just turned into a Bellhead. Oh, the irony.

Posted by Martin Geddes at 05:40 PM
Trackback Pings

TrackBack URL for this entry:
http://www.telepocalypse.net/cgi-sys/cgiwrap/mgeddes/MT/mt-tb.cgi/755.

Comments

This week I've heard diametrically opposed views on this. I think that we actually need two identities each (one for work, one for home). More at http://www.richard-marshall.com/wordpress/?p=21

Posted by: at July 2, 2006 07:55 PM
Please enter your comment below. Your comment will not appear immediately -- they all go for pre-approval by me because of the volume of spam I receive.







Remember personal info?