Comments on: let the dream die!

By: Tom

Tom — Tue, 06 Nov 2007 20:56:28 +0000

I think we’re probably closer than you think on the question of moving data. We can calculate the entropy of the piece of compressed data, for instance, and see how close it is to its maximally compressed (this only applies to lossless compression, of course). Similarly, we can take the bandwidth of a given slice of spectrum in a given medium and know how much data we can ever pass through it.
Now I’ll admit that important work is still being done — just yesterday I read about an Australian researcher’s development of a way to improve DSL efficiency. But my central point is that packet-switched networks offer such great advantages that it’s inconceivable that we’d move back to specialized signal-delivery systems rather than building new systems atop our present, unified, packet-based architecture (one that can, of course, have its capacity expanded as necessary).

By: Shun

Shun — Tue, 06 Nov 2007 19:09:11 +0000

Great analysis.
I only disagree on one point: “…and we’ve pretty much figured out the best ways to move those packets.”
I don’t think we’ll ever get to a point where we know the best way to move “packets”, if by packets, you mean information.
TCP/IP works, for now, because it’s open, ubiquitous, and built by nerds. Sorry, I don’t know why that last bit is important, but I do believe that it is.
Anyway, the companies are bound to fail. I mean, look at them: they’re companies. They need profits. They are alienating their support base. It’s bound to end badly.

By: Brendan Cullen

Brendan Cullen — Tue, 06 Nov 2007 12:31:12 +0000

“The present situation is as if the gas company gave customers specialized ovens that could only heat up prepared meals bought from the utility â and refused to connect the gas line to any other kind of oven.”
That might be the best analogy of the net-neutrality subject I’ve ever read.