So here's a thought --- my disconnected-as-continuum vs. binary argument really
starts to become important if you look far enough out. Since we're looking to
physics to define the edge cases, think about c --- light speed defines a maximum
signal propagation velocity, hence latency ~ distance. We *might* be able to push
things out over the long haul so that every location on the planet essentially has as
much bandwidth as anyone wants --- actually, I don't believe this will happen, but
let's stipulate it for the sake of argument --- but any interplanetary or bigger
network is going to have a fuzzier notion of connected vs. disconnected.
> Lucas Gonze wrote:
> > My money says that it is far easier to account for transient nodes in software
> My moneys says I'll build me a wearable when a megapixel 30 degree opening angle
> display is down to a kilobuck. (I'll probably won't even wait that long, and will
> settle for a crappy SVGA resolution, and a cellphone as wireless modem).
> > than to build fat always-on pipes to every square inch of the globe. Assuming
> Vacuum has a nice property: it propagates electromagnetic radiation. Physics
> doesn't forbid you having a TBps pipe if you're in line of sight to an optical
> router. Also, this universe is kind enough to allow orbits, which enables you
> to hang the skies with sufficient amount of hardware to dim the sunlight (no
> need to get that extreme, a few 100 microsats launched for a few megabucks
> will give you global coverage with quite a few kbps).
> > that developing nations will get always-on connectivity, when even Silicon
> > Valley doesn't have it w/out plenty of trouble, is a mistake.
> Right now every spot of the earth has got always-on connectivity, it's just
> it's not broadband, and most people can't afford it. Haul to LEO is getting
> cheaper year by year, though, and them sats are getting smaller and brighter.
This archive was generated by hypermail 2b29 : Sun Apr 29 2001 - 20:26:08 PDT