Re: Review of Vinge's Deepness in the Sky
Tue, 2 Mar 1999 14:39:34 -0800

Robin Hanson, <>, writes:
> So does Vinge present a plausible detailed picture? I'm not sure.
> Limits to software complexity were plausibly presented, and so I
> could buy the lack of AI or advanced automation. Though the story
> doesn't say so, I suppose complexity limits could also explain the
> life extension limits described. The failure to make substantial
> progress in physics seemed more arbitrary, though I suppose very
> subtle effects might remain hidden for millennia until the right
> clues were presented.

Vinge has suggested elsewhere that the Zones actually are limitations on computation. In that case we might expect that in the Beyond, some new kind of computational physics comes into play. Maybe it is possible to make a new kind of computer which solves NP complete problems efficiently. Those computers would be used to create AI and also to solve the problems needed for FTL travel. It would streamline programming and so get past the complexity bottleneck.

> More puzzling was the failure to achieve anything like nanotech.
> I suppose complexity limits could be behind this. In one case,
> a system with "a technology as high as Humankind ever attained"
> achieved something close to nanotech, and the dust our hero
> bought from them became a core element of all trader's starships,
> and the key to our hero's power. But I don't recall that system
> being noted for any other abilities to handle complexity. (It
> was particularly bad at life extension.)

I thought this "dust" was not that far ahead of our own capabilities with regard to micro electromechanical systems (MEMS). It was a fraction of a millimeter in size and had limited capabilities of sensing and manipulation, with moderate computing power. Such devices can be built using bulk manufacturing and don't necessarily represent a step towards nanotech.

> Perhaps most puzzling is the failure to use any significant
> fraction of the resources at each solar system. Human populations
> around a star are never more than "billions", and we see nothing
> like wholesale conversion of asteroids and comets. "Sooner or later
> [each system] ossified and politics carried it into a fall."

We see a system where they have habitats in the asteroids and they've terraformed a couple of moons, in addition to living on a planet. Beyond that they might have had to start taking gas giants apart, which would take a long time. It could be that after a few cycles people have adopted a fatalistic attitude like Niven and Pournelle's Moties. They know that civilization is only going to last a few thousand years, and long term projects don't make sense given those constraints.

The one fall we see in detail seems to have been due to software failures, overly complex protocols whose accumulated bugs are driving the whole system to a halt. It seems that Vinge is relying heavily on this buggy-software idea to explain the way his universe works.

> The frustrating thing about using science fiction to think
> about these issues is not knowing whether the author thought
> they had good reasons to expect things described, whether they
> were just choices to make the story easier to tell, or whether
> the author just didn't even notice them. I suspect one big
> problem is that Vinge doesn't really believe in these limits.

He's got to prevent Singularities for dramatic purposes. Given that we don't have Singularities, what stops them? Software complexity may not be a bad choice as a barrier.