Monday, May 30, 2016

Weathery Problems

Alien civilizations who push the knowledge of technology a bit farther that we have eventually reach what we term 'asymptotic technology'. This is the ultimate state, when the alien civilization knows just about all that science can offer, and has the engineering knowledge to boot about how to turn it into useful items for the use of the members of their civilization. It is akin to omniscience, but it has limits, imposed not by the characteristics of the alien civilization, nor by the gaps or errors in their knowledge, but by the very nature of the universe.

Asymptotic technology is universal, meaning that every alien civilization that ever existed will reach the exact same body of knowledge, as it does not depend on any details of the civilization, but upon the details of science and engineering. This also means that exactly the same limits will impact each and every alien civilization, not counting the ones who become extinct early in their progression or who run into problems such as Malthusian idiocracy. Those who are successful get to exactly the same point in technology, and then according to the principle of technological determinism, their societies will conform to the same technology and have a great deal in common with every other alien civilization that gets to the same peak.

The same limits to asymptotic technology hold no matter when the alien civilization encounters them. They are hard limits, and if an alien civilization gets far ahead in one subspecialty of science and hits a limit, it is going to be stuck with that limit until the end of its existence. Asymptotic technology does have an order to it, as some advances require technology from other branches in order to proceed past some thresholds, but these are not so restrictive that they insist that each subspecialty only can arrive in some particular sequence. It is the limits which are universal and there is some flexibility in which ones are bumped into first.

The same technology limits impact our work here, meaning both the work that Earth's scientists do, and also the research we are attempting to pull together on alien civilizations. As an example of the first, we use the old, old example of the weather. No one in touch with modern media doubts that weather forecasters have a very hard problem predicting weather for more than a few days, if that. Some generic climatological averages exist for long term predictions, describing the year and perhaps some correlations between different aspects of a year's climate, but to predict rain thirty days in advance simply is not even attempted. The reason for this is thought to be well-known.

We are nowhere near the asymptotic limits of computational power, as evidenced by what we call Moore's law, which says something to the effect that computational power is growing at an exponential rate, or at least used to be. So if we jump ahead to the future, can we expect that when computational power grows ten or a hundred times as great as it is now, we will be able to predict rain thirty days in advance? Predictive power will certainly improve, but will it be incremental, a few percent improvement in short term forecasts, or an order of magnitude in the accuracy of predictions way out in time from the current time? The former. Because computational power is not the only obstacle to weather calculations. Data is another one. Without a corresponding improvement in data, computational power applied to existing data would accomplish nothing at all. Data means having a compendium of all those variables that affect local weather, meaning temperature, humidity, pressure, velocity, cloud cover, composition, and perhaps others everywhere in the atmosphere from the surface through the exosphere, in three dimensions. With all that data, plus data on solar impact, orbital variables, surface temperatures, and likely oceanic data as well, computational models would have a better starting point. If models became improved as well, so the computational fluid dynamics needed to make the computations were as sound as could be, they might make the best possible effort toward computing rain thirty days out.

There may be a snag, relating to scale. Just exactly how precise in three dimensions does this data have to be? We don't have a clue as to whether a data collection system would have to collect data every kilometer horizontally and every hundred meters vertically, or whether it might need twice or four times that much, or perhaps more? This is the sensitivity problem. Do the predictions work if they are averaged over a kilometer, or only over a hundred meters?

For lack of a better word, let's call a three-dimensional problem with heavy computational requirements and very heavy data requirements a 'weathery' problem, i.e., a problem like the weather. There are several others, and they impact the scientific work that is needed to provide some answers to questions posed by studying alien civilizations. That means that there will be some questions without answers for a very long time.

One such problem relates to the formation of stars from gas clouds. It is certainly possible for scientists to build models of one or two-dimensional star formation, but gas clouds are non-uniform. Even if all the brilliance in the world was focused on the problems of nuclear fusion under immense pressure, and the results were astonishingly promising, there is still the problem of analyzing any particular star, or stars with three-dimensional clouds originating them, which is a weathery problem.

Suppose we understand that the presence of sufficient uranium and thorium resources on a rocky planet are critical to an alien civilization getting to a final level of technology. Determining the separation of uranium ores on the upper crust of some arbitrary planet is a weathery problem, and developing conditions for it to happen is a very difficult assignment. Assessing if fusion is possible, at least in the form we are experimenting with it on the largest experiments so far, is a weathery problem. If it were not, it would be possible to calculate just what would work; but it is and so experiments have to be conducted.

Figuring out how the two disks evolve, the galactic disk and a planetary disk, are weathery problems. Even N-body problems can be called weathery problems, and examples of multi-planetary mutual perturbations and globular clusters are included. Problems in what we call soft sciences, psychology, economics, sociology and even neurology are all computationally difficult, if there were already any computational models, which there are not in most cases. Each of these weathery problems will not easily be solved, not with the next ten years of computer development nor the next ten years of observations or data collection. In each case, some sage speculation is needed. That is the best that is possible, so the study of alien civilizations is not going to look like atomic physics any time soon.

No comments:

Post a Comment