Iterative Solutions, or why planning 10,000 years ahead is a bad idea.
06.04.08 09:16 Filed in: Philosophy
| Narratives
| Operation
of Complex Systems | Change
Management | Configuration
Management
I was musing about the failure of the U.S.
government to come to terms with its
responsibility and obligation to fulfill on the
promise of long term high level radioactive waste
storage with my friends and mentors the other
day. I was struck by the concern that seems to be
rampant among the public, the responsible
government officials, and the scientific and
technical effort associated with the Yucca
Mountain project that we somehow provide a long
term high level radwaste storage solution that
will be viable for 10,000 years.
Now maybe it's just me, but doesn't that strike you as ludicrous? We've barely been around that long as a species, at least as far as I've heard, and we have the arrogance to try to project a solution ten thousand years into to the future. How absurd, I trust you are saying about now. To declare that your assumptions regarding long term radwaste storage will be valid ten thousand years into the future is not only foolish, self-serving, and irresponsible, it does nothing short of guarantee the failure of the Yucca mountain effort.
But this is not meant to be a railing diatribe against inane government policy, even if it may in fact be inane. (Currently the EPA is proposing a regulation which will extend exposure limits to one million, that's right, 1x10E6, years. Guffaw!) Rather, I started speculating about the probability that a logical argument could be built against trying to finalize a solution that was expected to endure even more than 250 years. What follows here is not that argument, but some building blocks for the argument, to provoke a little conversation and some exploration of possibilities that might be effective in resolving the 'solution' enigma.
Point A: History. At the time of the Roman empire, no one could have conceived that trans-global transportation, moving a person from say, Rome, to say, the yet unbuilt Chicago, could take less than 15 hours, as it does now, 1800 years later. You wanted to go somewhere, you marched. You wanted to go somewhere fast, you rode a horse or chariot. Humans could not see as far as the Boeing 777 burning refined petroleum landing on concrete runways three feet thick. In the 1600s, no one could have conceived that you could move tens of thousands of tons of product hundreds of miles for a few dollars a ton. They could not see as far as the GE Electromotive Division locomotive and rail infrastructure that connects freight with consumers nation wide on every continent. In 1900, the possibility that a man might walk on the Moon was science-fiction at best. But 69 years later, the Eagle landed, and what was inconceivable in the year 200, unimaginable in 1600, and nothing more than fantasy in 1900 became so commonplace as to be ignored in the prime time news hour. Even 40 years ago, no one would have predicted that we could cultivate and harvest grapes in Chile, pick and pack them, fly them to the US on a 747, unload, ship, store, and distribute them to your local grocery store in time measured in hours, and still sell them for seventy-nine cents a pound.
Each of these advances required an iterative approach - we had to try and sometimes fail at several intermediate points in order to accomplish the goal that we now appreciate as commonplace.
Lesson learned: Interim situations are necessary, even required, for expanding our capacity to act in the future. We should approach something as critically serious as radwaste storage in a series of solutions.
Point B: More history lessons. History demonstrates that the cycle time for improving the effectiveness of solutions decreases with each iteration, that is, each time we effect a solution to a problem, the solution itself (Law of Unintended Consequences?) reveals breakdowns that the solution did not address. I don't know what the equivalent of Moore's law is for solution improvement, but I'll speculate that it is measured in years or tens of years, four or five orders of magnitude less than the Yucca mountain solution.
Lesson learned: Identifying breakdowns fast and often moves you quickly through the learning curve of unanticipated issues to a successful implementation.
Point C: As Will Rogers so gracefully said, "It's not what we don't know that causes us trouble, it's what we know that ain't so!" Human beings do not have much capacity for seeing the future. Meteorologists are proven wrong routinely within 24 hours, investors leap from windows to punctuate their skill in predicting the market, and the only people who consistently make money on horses are the bookies. And these are situations that are a full five orders of magnitude closer to the present in time than the Yucca Mountain solution. When we assume, or make an assumption about how things are going to be in the future, we necessarily must act as if that assumption is true. So as a consequence, we plan as if something is true, when for all intents and purposes, we can be almost certain that it is not.
Lesson learned: Our assumptions about the future that provide foundation for our design are only as valid as our ability to predict the conditions under which the assumptions must remain valid, which we have seen historically is not very accurate. Shortening the range that our assumptions must survive makes our design more robust.
Point D: Proponents for the one million year requirement make the assumption that the site will not be occupied, monitored, and improved during that one million years. As far as I can tell, no one has suggested such a thing as part of the design, construction, operation, and maintenance of Yucca Mountain. Historically, we identify failures in our design by usage and observation, not by guessing at how long we can stay away without anything going wrong.
Lesson learned: Continuous observation allows us to anticipate when and where breakdowns will occur, and modify the design to accommodate the new learning.
Point E: Technology is evolving at an ever increasing rate. One only need look at the last hundred years to realize that what we believe are the limits of technology today will be meaningless 10 years from now. When I was a boy, traveling from Chicago to Michigan for summer vacation meant 2 hours driving through the black hole of northern Indiana steel country. The sun was a vaguely orange disk through the cloud of black metallic tasting smoke from the steel mills that hung for 50 miles along the lake shore. People were so gripped by the story of steel that they couldn't conceive of life without the black cloud, and gritty fall-out that coated everything. That was the 60's. Twenty years later, the clouds were gone, because we had invented technology to eliminate them. Today, the steel mills are essentially gone, because their technology is no longer viable.
History has shown that we are so unable to predict the technological future that we can't begin to imagine how nuclear radwaste storage will evolve over the next ten to twenty years. Prognosticator that I am, I'll go out on a limb and suggest that 250 years from now, radwaste won't be an issue, because there won't be any, meaning we will have invented technology that eliminates the production of radioactive waste. There, I've said it. Remember you heard it here first.
Lesson learned: Time doesn't stand still - evolution in technology continues at a breakneck pace. If this is a technical issue now, expect that it will be resolved in 250 years.
Point F: When we seek the perfect solution, nothing happens. No progress is made as we argue about the perfect and final solution. We are arguing about assumptions that we can no more prove than we can ignore the laws of physics. And nothing gets done.
Lesson learned: Do something, even if it is not complete. As Patton said, "A good plan violently executed now is better than a perfect plan executed next week!"
There are probably more points to be made, but this is enough to get started.
Where do we go from here?
First, abandon the existing plan, and shoot anybody who suggests that more than 250 years is necessary. Well, don't actually shoot them. Banish them from the halls of rational thought, because they don't belong there.
Two, develop the 250 year solution. Two hundred and fifty years is enough time to evolve a better solution, and put it in place.
Three, take all that money you saved by not developing the perfect solution, and create an investment account that pays for the execution of the 250 year solution, and for developing the next solution. I'm confident that we won't need more than one or two "250 year" radwaste storage solutions. If we do, send your complaints to this address. See what I mean?
Four, tell your Senators and Representatives what you think. You can quote me if you like, but your words will probably come from the heart, and they need to hear from you, because they aren't thinking clearly.
I'm guessing that there is a paper that needs writing in all this. Any volunteers?
I would be interested in hearing your comments and feedback.
Now maybe it's just me, but doesn't that strike you as ludicrous? We've barely been around that long as a species, at least as far as I've heard, and we have the arrogance to try to project a solution ten thousand years into to the future. How absurd, I trust you are saying about now. To declare that your assumptions regarding long term radwaste storage will be valid ten thousand years into the future is not only foolish, self-serving, and irresponsible, it does nothing short of guarantee the failure of the Yucca mountain effort.
But this is not meant to be a railing diatribe against inane government policy, even if it may in fact be inane. (Currently the EPA is proposing a regulation which will extend exposure limits to one million, that's right, 1x10E6, years. Guffaw!) Rather, I started speculating about the probability that a logical argument could be built against trying to finalize a solution that was expected to endure even more than 250 years. What follows here is not that argument, but some building blocks for the argument, to provoke a little conversation and some exploration of possibilities that might be effective in resolving the 'solution' enigma.
Point A: History. At the time of the Roman empire, no one could have conceived that trans-global transportation, moving a person from say, Rome, to say, the yet unbuilt Chicago, could take less than 15 hours, as it does now, 1800 years later. You wanted to go somewhere, you marched. You wanted to go somewhere fast, you rode a horse or chariot. Humans could not see as far as the Boeing 777 burning refined petroleum landing on concrete runways three feet thick. In the 1600s, no one could have conceived that you could move tens of thousands of tons of product hundreds of miles for a few dollars a ton. They could not see as far as the GE Electromotive Division locomotive and rail infrastructure that connects freight with consumers nation wide on every continent. In 1900, the possibility that a man might walk on the Moon was science-fiction at best. But 69 years later, the Eagle landed, and what was inconceivable in the year 200, unimaginable in 1600, and nothing more than fantasy in 1900 became so commonplace as to be ignored in the prime time news hour. Even 40 years ago, no one would have predicted that we could cultivate and harvest grapes in Chile, pick and pack them, fly them to the US on a 747, unload, ship, store, and distribute them to your local grocery store in time measured in hours, and still sell them for seventy-nine cents a pound.
Each of these advances required an iterative approach - we had to try and sometimes fail at several intermediate points in order to accomplish the goal that we now appreciate as commonplace.
Lesson learned: Interim situations are necessary, even required, for expanding our capacity to act in the future. We should approach something as critically serious as radwaste storage in a series of solutions.
Point B: More history lessons. History demonstrates that the cycle time for improving the effectiveness of solutions decreases with each iteration, that is, each time we effect a solution to a problem, the solution itself (Law of Unintended Consequences?) reveals breakdowns that the solution did not address. I don't know what the equivalent of Moore's law is for solution improvement, but I'll speculate that it is measured in years or tens of years, four or five orders of magnitude less than the Yucca mountain solution.
Lesson learned: Identifying breakdowns fast and often moves you quickly through the learning curve of unanticipated issues to a successful implementation.
Point C: As Will Rogers so gracefully said, "It's not what we don't know that causes us trouble, it's what we know that ain't so!" Human beings do not have much capacity for seeing the future. Meteorologists are proven wrong routinely within 24 hours, investors leap from windows to punctuate their skill in predicting the market, and the only people who consistently make money on horses are the bookies. And these are situations that are a full five orders of magnitude closer to the present in time than the Yucca Mountain solution. When we assume, or make an assumption about how things are going to be in the future, we necessarily must act as if that assumption is true. So as a consequence, we plan as if something is true, when for all intents and purposes, we can be almost certain that it is not.
Lesson learned: Our assumptions about the future that provide foundation for our design are only as valid as our ability to predict the conditions under which the assumptions must remain valid, which we have seen historically is not very accurate. Shortening the range that our assumptions must survive makes our design more robust.
Point D: Proponents for the one million year requirement make the assumption that the site will not be occupied, monitored, and improved during that one million years. As far as I can tell, no one has suggested such a thing as part of the design, construction, operation, and maintenance of Yucca Mountain. Historically, we identify failures in our design by usage and observation, not by guessing at how long we can stay away without anything going wrong.
Lesson learned: Continuous observation allows us to anticipate when and where breakdowns will occur, and modify the design to accommodate the new learning.
Point E: Technology is evolving at an ever increasing rate. One only need look at the last hundred years to realize that what we believe are the limits of technology today will be meaningless 10 years from now. When I was a boy, traveling from Chicago to Michigan for summer vacation meant 2 hours driving through the black hole of northern Indiana steel country. The sun was a vaguely orange disk through the cloud of black metallic tasting smoke from the steel mills that hung for 50 miles along the lake shore. People were so gripped by the story of steel that they couldn't conceive of life without the black cloud, and gritty fall-out that coated everything. That was the 60's. Twenty years later, the clouds were gone, because we had invented technology to eliminate them. Today, the steel mills are essentially gone, because their technology is no longer viable.
History has shown that we are so unable to predict the technological future that we can't begin to imagine how nuclear radwaste storage will evolve over the next ten to twenty years. Prognosticator that I am, I'll go out on a limb and suggest that 250 years from now, radwaste won't be an issue, because there won't be any, meaning we will have invented technology that eliminates the production of radioactive waste. There, I've said it. Remember you heard it here first.
Lesson learned: Time doesn't stand still - evolution in technology continues at a breakneck pace. If this is a technical issue now, expect that it will be resolved in 250 years.
Point F: When we seek the perfect solution, nothing happens. No progress is made as we argue about the perfect and final solution. We are arguing about assumptions that we can no more prove than we can ignore the laws of physics. And nothing gets done.
Lesson learned: Do something, even if it is not complete. As Patton said, "A good plan violently executed now is better than a perfect plan executed next week!"
There are probably more points to be made, but this is enough to get started.
Where do we go from here?
First, abandon the existing plan, and shoot anybody who suggests that more than 250 years is necessary. Well, don't actually shoot them. Banish them from the halls of rational thought, because they don't belong there.
Two, develop the 250 year solution. Two hundred and fifty years is enough time to evolve a better solution, and put it in place.
Three, take all that money you saved by not developing the perfect solution, and create an investment account that pays for the execution of the 250 year solution, and for developing the next solution. I'm confident that we won't need more than one or two "250 year" radwaste storage solutions. If we do, send your complaints to this address. See what I mean?
Four, tell your Senators and Representatives what you think. You can quote me if you like, but your words will probably come from the heart, and they need to hear from you, because they aren't thinking clearly.
I'm guessing that there is a paper that needs writing in all this. Any volunteers?
I would be interested in hearing your comments and feedback.
|