Its doesn't matter if the person's movement lags by 200ms, even if you were in the same room you probably would not notice. For meeting with someone over long distance really doesn't matter.
Typically in VR applications you buffer the last 3 packets just to make up for lag and frame jitter. It doesn't matter how much the lag is, it's much more important that the user doesn't experience it.
I have been writing applications that do VR meeting for 17 years, so I do know a little about the subject.
Even the latency on the smartphone display-loop is terrible (try moving your finger back and forth), mice were better; we're just used to it. MS experiments show that a perceptual transformation occurs at around 10-15ms. The same figure pops up in VR lit.; also in electronic music instruments (where it has been solved). [\tangent Perceptually-instant smartphone interactions are a potential revolution.]
And since it affects us when interacting immediately, it also affects us when interacting remotely.
Wow, the VR problem is latency? Have you ever had a real-time conversation with anyone outside your hemisphere? If 15ms is 15,000 feet then 1/0.015 = 66 times that should be the speed of light, i.e. 300,000 km/s. Does that (66 x 15,000 ft = 300,000 km) seem even vaguely in the right ballpark to you?
You're off by a factor of 1000. It's 15 microseconds, not milliseconds. And incidentally, this might be a good excuse to go travel outside the country and meet people from elsewhere on the planet... you never know what you might learn about the world around you.
I assume the 15ms was chosen for the 60fps that VR needs to run at to be considered reasonable. If so, having a local machine handle redraws would be reasonable and requiring a 30fps update on all "agents" and other world elements would probably be pretty unnoticeable by most people.
Anyone can notice the total latency in even a Skype conversation.
The part that's not correct is 15ms ~ 15,000 feet @ c:
15ns ~ 15 ft
15us ~ 15,000 ft
15ms ~ 15,000,000 ft
Some other fun figures: https://www.wolframalpha.com/input/?i=earth+circumference
Light travel time t in vacuum from t = x/c: | 134 ms (milliseconds)
Light travel time t in an optical fiber t = 1.48x/c: | 198 ms (milliseconds)
If it had been a typo or just an arithmetic error, I wouldn't have cared (or even noticed). But you also completely confidently declared that the problem had been identified to be c and that the boundary for VR is "city limits". It means that either [1] you have never had a real-time conversation with anyone much farther away to help you sanity-check this result (in which case it's an innocent mistake, and like I said, it's a good excuse to also go out and see the world), or (or possibly "and") that [2] you did not even care to try to think about this topic for 2 seconds before writing about it -- not when you first heard about it wherever you did, nor when you wrote about it here to teach it to other.
Whatever the larger issue, the silly math error is 100% beside the point... you could've written light-years for all anyone cared, and forgotten about it the day after. However, the underlying issue is not the kind of problem to forget about the next day. Lots of alarm bells should be going off in your mind that you need to actively do something to correct whatever the root problem is (which I was trying to help you identify), because otherwise you'll keep actively making the world (and HN) a worse place with this kind of misinformation, and that is something I'm tacitly assuming you don't want to do.
So: please, please, please ensure put 1-2 seconds of thought into sanity checking whatever you write in the future -- especially it's too painful to at least qualify what you write with an "I think" so that readers realize you're deriving your own conclusion and not propagating an established fact. We have enough problems in this world with the spread of misinformation that hinges on subtle nuances that we don't also need to put up with problems that should be pretty easy to spot from light-years away.