I can understand why it made people uncomfortable—there was a certain intimacy to it. But, conversations are often healthier when the participants are forced to be a bit vulnerable.
The nice thing about Google Wave (other than the instant typing) was how you could seamlessly move between talking and collaborating.
Haven't tried or seen much of Microsoft Loop.
We were very young men, and he was coaching me in some situation with a girl.
He died soon after of CF. RIP nion
Ntalk is my second favorite way of communicating after in-person face-to-face talking.
/MagPie
This is that feature in xmpp, but it didn't send as you typed, it batched up typing and sent just one packet a second (or so) with timed keystrokes in the packet so it could play back a simulation of you typing; thus giving the illusion of live typing while not swamping the network.
There's some demo gifs over on https://www.realjabber.org/ (the more recent website had flash demos only and has since been shut down...this is all a bit delapidated)
IMO, this completely defeats the point.
The advantage of live typing is that I can see what someone is typing sooner, before they're even done, and start considering my response. I want to see their message, or partial message, as soon as possible.
I don't want to see some movie of all of their keystrokes and corrections. It's totally fine if it comes with the territory, but only in the name of reducing latency.
That isn't to say I'm against batching the keystrokes to conserve bandwidth, but don't fake it!
What isn't clear to me from the demo, however, is what happens if the typist uses backspace or pauses while typing. I wouldn't want those things to get transmitted if they happened within the span of a single transmission.
In other words, it's okay to display letter-by-letter, similar to video game dialog, but they don't need to keep the original timing information.