Re: Review of Vinge's Deepness in the Sky

Michael S. Lorrey (
Wed, 14 Apr 1999 13:01:48 -0400

Robin Hanson wrote:

> On 4/13/99, Wei Dai wrote:
> >>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."
> >
> >I think the central assumption Vinge used here is that no stable forms of
> >social organization are possible in the Slow Zone, ... Even the most
> >stable organizational forms become increasingly unstable as a civilization
> >develops technologically and economically, eventually suffering a
> >catastrophic collapse. The Slow Zone governments must try to delay this
> >collapse by restricting economic growth and research ...
> This theory needs to be augmented to explain why the "no organization"
> organization fails. Across the thousands of stars there is a weak
> organization via trade that doesn't seem to threaten the total human
> region. So why couldn't the same weak ties work within a star's system?
> Why couldn't a solar system fragment into thousands of places each of
> which was organized internally, but where ties between places were weak?

Purportedly, the vast space between the stars prevents opressive governments from extending their reach indefinitely, so the crashes are limited to the star system. Vinge speaks of two causes for falls:

  1. development of a total surveillance society/government
  2. if the first is avoided, eventually, the software infrastructure becomes so layered and bug ridden that it collapses under its own weight of complexity.

The second cause does not reach beyond a star system because software controlled systems do not extend beyond the star system, so when the software fails, the failure is limited to that star system.

I also wonder at this, why the Queng Ho do not develop and sell an all purpose single layer software technology which does not need further development, interpretation, or modification. If they did this, they could avoid the second cause. It should not be too hard to develop simple intelligent agents which routinely scour code for bugs, debugging as they go. We already have them for many types of compilers, even for web pages.

> >>These falls are very severe, often requiring re colonization from
> >>the outside, and otherwise seem to require rebuilding from
> >>scratch. This is much more severe than the fall of the Roman
> >>Empire, for example. ...

Imagine that much of a sytem's population lives in orbiting colonies or on airless moons. These populations' entire air, water, and food processing and distrupution systems are fully automated. When these systems fail, the people starve, suffocate, and die of thirst. When a planetary population lives in hive cities, it is much of the same, and when the food supply is automated hydroponics, or is based on seeds that have suicide genes like the Monsanto IP protection gene, then starvation on planets will happen.

> >
> >We also see biological weapons, which should be sufficient.
> I'm not yet persuaded of that. If bio weapons can kill all the
> humans in a star system, why doesn't it spread to other star
> systems via the trading ships? If trading ships can block
> the spread, why can't places within a star's system use the
> same approach?

Presumably because a bioweapon that severe will not allow a starship's crew to survive the voyage, so the crew will not be there to respond to hails by the port crews at the end of the trip, so the ship will be destroyed or allowed to cruise on through the system and back into the depths of space.

This is similar to the spread of the Black Plague. while ships helped spread it to other ports within sufficient distance such that only a small percent of the crew succumbed during the voyage, the port authorities at the end of the voyage would often refuse docking to a ship they feared was a plague carrier.

Mike Lorrey