There is another possibility we should consider: a society that has retained the capacity to operate and maintain technology, but not create or understand it. We seem to be increasingly moving toward such an era -- I don't have numbers, but I think the ratio of people who understand technology versus those who use them has been gradually dropping over the past fifty years, mainly because more and more people are using technology and the technology is becoming automated. A scientifically illiterate society with giant mining machines is not entirely outside the realm of possibility.

I've been working with a Kiwi for a bit now. He grew up on the only 'computer' in NZ back in the 70's, a punch card machine that he had to actually fight (with real fists he says) to use time on. This guy has seen the digital revolution up close.

There was an article on HN a while back that mentioned the #1 programming language being taught today was Python. I told the Kiwi this and he just sighed. Python is a 'batteries included' type language. You don't have to really worry about any memory issues, any special loops that will go to infinity, none of the real stuff. Yes, it teaches you the 'idea' of programming, but programming is much more than an idea, it is a set of instructions you send to an electrical machine to preform a task. If you take away the entirety of the actual physical world, you have done the kids a disservice. Programming, real programming, is not supposed to be easy.

I'll mention the Kiwi was here in the US on a work holiday. I was trying to get a chip working (yes, with actual registers) and spent ~6 months on it. The Kiwi, I kid you not, spent about 3 hours. The guy really did know his stuff.

> none of the real stuff

It's really good to know how things works underneath, but insisting that constantly messing with low-level details is "real" programming (and implying high-level details aren't "real") is weird. Both aspects - high and low-level - are important.

Beginners are taught high-level languages because it's easy to start with - one could create something useful without learning lots of things. If they care about their craft, eventually they will learn all the gory details of the stack underneath down to bare metal - an occasions (bugs, weird behaviors or just curiosity) where, say, Python programmer is invited to go and peek under the hood of their interpreter or libc or kernel or even hardware happens every now and then.

Unless all that pythonista's doing is bashing some boring CRUDs with Django. Then, a knowledge of low-level stuff seems almost unnecessary. But, hey, someone has to do that kind of programming, too.

>We seem to be increasingly moving toward such an era

You should visit a Maker Faire. I can't think of anything that restores my faith in humanity quite as much as a day at one.

I very much doubt that people are less sophisticated about science and technology than any time before. Quite a contrary, the education system has given us laymen discussions in which scientific literature is referred regularly (HN, many popular subreddits, even discussion threads in my Facebook stream)

Due to a lack of man hours available to teachers, we make up for breadth instead of depth. This is a good thing, as we do not know going forward what we should teach in depth to children. Teaching kids about cam shafts and carburators would be a fools errand, but you only knew that about 5 years ago. However, depth is needed in many areas, we just have to trust people to figure that out on their own, as they are more mobile than an educational system. An example is bit wise operations in computers. Many don't have a clue as to how to use a register in a chip, and we thought we'd never have to again. But encrypting, securely, relies entirely on this and trying to poke holes in it. Now, again, we need the depth there to maintain encryption as a viable idea.

I don't have numbers either, but I think that we are continuing to specialize. We have an ever increasing portion of the population understand technology, but any given technology can be used by many and understood by few.

Idiocracy (http://en.wikipedia.org/wiki/Idiocracy) is also based on that idea.

Pump Six is a great story about this kind of future.

"You need electricity for that"

You're extremely close to the correct answer there. This topic suffers from the same fixation as digital archivists, how do we abandon something for 10K years yet find it "working" in 10K years? Oh that's easy, don't abandon it.

So you need a way to maintain the site, with some level of continuity. Well, I'm guessing in the middle of a desert, watermelon farms and recreational waterparks and aquaculture aren't going to work. How about the worlds largest solar power plant?

That would provide a reason for staffing and monitoring the site and a reason not to dig it up, or at least not to dig far enough down to get the bad stuff. If you build it "tough" and "reliable" there's no reason to think it would ever be decommissioned. You don't even technically need to run a profit, just need to generate enough KWh that it'll always be able to support, however minimally, some kind of onsite staff to keep it under control and keep things "civil".

Perhaps put a giant military base on it. With the idea that when civilizations collapse, the army seems to be the last entity to go. Of course that makes it a strategic target, but if the bad stuff is buried 3 miles down, or maybe nearby but not underneath, its still OK.

The other fixation is the 10K years. Trust me, if you run the math, its not magically turned into organic applesauce at year 10001... Much as the radiation exponentially decays, so does the need to "protect" and "document" the site. Somebody cracking it open in 500 years is a lot bigger problem than 5000 years or 50000 years, so we need to focus on 500 yrs problems a lot more than 50000000 years.

By far, the biggest danger the site poses is 500 yrs from now someone with a bit of ambition is going to point out that with centuries of "used" fuel with burnup ratios like 1%, that means 99% is unburned and with a bit of refining, we could run the worlds reactors for 99% of a couple centuries. Of course refining has some environmental risks, and the only thing worse than a local getting the bright idea, is someone "far away and upwind" being owed a favor and not particularly caring what happens downwind. The most likely environmental disaster failure mode isn't ignorant savages from the intro scene of the 2001 movie trying to make a hammer out of nuke waste, but some kind of economic collapse in a couple decades and Saudi Arabia is willing to trade U235 for their atomic powered desal plants in exchange for crude oil. Or maybe trading with Canada to get food, whatever.

Finally it doesn't really matter anyway. The fixation on nuke waste is its only dangerous for "awhile". We have much larger stockpiles of ore refining waste, coal ash waste, organic chemical waste, all of which lasts forever, and no one cares about that, although in practice its a much bigger problem.