Here is the paper, via google scholar.

Here is my stab at a solution (already posted at ffnet):

First Harry tells V. that Dementors are death, Patronuses work by not thinking about death and the true Patronus works by using a diferent mindstate which V. probably cannot attain (without specifics). Second Harry states that as long as Dementors are around every person including V have in each moment a small but finite probability to be kissed by one. Over an indefinite timeframe the aggregate probaility that V. is kissed approaches one. How this would interact with V's Horkruxes is unclear but he may easily suffer a fate worse than death. Therfore he should keep Harry around at least until the dementors are dealt with.

Then he points out that given what he knows about the ambiguity of prophecies the prophecy V. heard has probably not clearly identified that Harry and not V. is the threat. Thus V. killing Harry might easily doom the world. This is especially likely as V. is not bound by the vow. Thus V. should keep Harry around to guard against his own mistakes and probably take a similar vow. He himself may offer more vows to further Vs goals in exchange for V. vowing to further Harry's goals and so on. This should be beneficial for even a purely selfish V. who wants the world to survive.

In case V. is not convinced by his above offer of cooperation Harry uses the time they are talking to prepare for an attack on V. and the Death Eaters using partial transfiguration: Thinking about venues for attack he first thinks about transfiguring an invisible nanoweapon such as a monofilament knife to decapitate the death eaters. Though he quickly realizes that that will not work since no known material including carbonanotubes is stiff enough to form an invisible blade of several meters length. Independently acting nanobots are out too, because he lacks time and knoledge to design one let alone test them for safety and efficiency. Then he realizes he does not need them, because partial transfiguration can do everything a nanobot could and even more.

He points his wand to a patch of skin on his leg and starts to transfigure the stratum corneum. An invisible bundle of carbonanotubes extends from his skin to the ground branches out to each death eater running up their robes and into their necks. (They do not feel this, since the bundle of tubes has a crossection of only 50 nm. Pain or touch receptors would not pick that up.) Another branch extends to the Dark Lord, but Harry does not dare touch him with his construct fearing the resonance. Instead he builds a small tower form the ground using carbonanotubes in a pattern resembling the Eiffel Tower extending right into the muzzle of his gun (Beneath the moonlight glints a tiny fragment of silver, a fraction of a line...). He seals the muzzle with a thin sheet of carbonanotubes and fills the barrel with nitroglycerine contained by a second thin sheet of carbonanotubes just before the bullet. All of this is very low volume and quickly transfigured.

If the Dark Lord refuses cooperation he snaps his fingers and immeadetly extends the tube in each of the death eaters neck to severe the brainstem from the spinal cord, the language center from the brain (to prevent wordless, wandless magic) and the neck from the body (black robes, falling). To make sure that everything is properly seperated he turns his entire construct (except for the part in Vs gun) into pressurised air (...blood spills out in litres,...). Now the Dark Lord either surenders or fires his gun. ...and Harry screams a word: "rennervate" and points at Hermione to wake her up. Hermione stunns V. Even if V. fired he should not die immeadetly except if part of the gun passed through his brain. Hermione transfigures V. into a small stone to prevent him from dying and thus from coming back. Afterwards they transfigure the Death Eaters for eventual revival.

I wrote multiple redundant plans, because I genuinely think Harry should be able to convince V. to cooperate for purely selfish reasons. But even if V. is not only rational and selfish but "For the Evulz" Evil and thus refuses, the transfiguration attack should secure Harrys victory.

I was going to say "doctor's don't have the option of not picking the diagnosis", but that's actually not true; they just don't have the option of not picking a treatment. I've had plenty of patients who were "symptom X not yet diagnosed" and the treatment is basically supportive, "don't let them die and try to notice if they get worse, while we figure this out." I suspect that often it never gets figured out; the patient gets better and they go home. (Less so in the ICU, because it's higher stakes and there's more of an attitude of "do ALL the tests!")

Survey taken!

Concerning the mental health questions, how do you weight self diagnosed and diagnosed by psychiatrist? Do you think, given the Less Wrong demographic self diagnosis is less or more reliable (intuitively I would tend to more). How should cases like myself answer - diagnosed with asperger by psychiatrist1, two years later diagnosed with ADHD but not asperger by psychiatrist2, several month later diagnosed as neither asperger nor ADHD by psychiatrist3?

'cause nothing they know is particularly useful for future generations.

Knowing how to make hydraulic cement isn't useful?

I am quite certain that the 10 000 top engineers and scientists know quite a few things that would be very useful for future generations. Since I am not in that number, and since I am only one person, I do not know what those things would be, but I estimate a high probability that they exist.

But can you make hydraulic cement underwater? I was under the impression that you needed fire to make it.

I'm not sure. The Wikipedia article mentions that the ancient Romans used a mixture of volcanic ash and crushed lime, and you certainly do get underwater volcanoes, so the ash should be available... there are probably industrial processes now, but just mixing volcanic ash with the right sort of mud and getting something that hardens if you leave it for a day or two sounds usable underwater to me.

You could always argue that we are both not creative / intelligent enough to find a solution and that this is not indicative that a whole society would not find a solution. And this argument may well be correct.

Given an expectation of how hard it is to solve the problem if it can be solved, inability to solve it with given effort produces corresponding evidence of impossibility of solving the problem. Not responding to inability to solve the problem amounts to actually expecting the problem to be very hard. If I don't expect that, I would be wrong in suggesting that inability to solve the problem is not evidence for impossibility of solving it.

Another framing is to generalize "inability to solve the problem" upon the conclusion of the project, to a situation where the expectation that the problem can be solved eventually is reduced. Correspondingly, generalize "ability to solve the problem" with expectation having gone up upon the project's conclusion. This way, it's clear by conservation of expected evidence that you can't expect that the estimate for the probability that the problem is solvable will go in a particular direction upon the conclusion of the project. Either the expectation will go up (and so the project produces evidence of the possibility of eventually solving the problem), or else it must go down (and so you gain that elusive evidence of the negative).

Can you imagine a way a group of quadriplegics (imho a good aproximation of a stranded dolphin with a human brain - except that their skin does not dry out - ) could fell a tree with stone tools? And delimb it? And bring it to the construction site? And erect it as a pillar?

Sure, depending on what you are thinking about as the reference procedure of, say, chopping down a tree when using hands. Dolphins with hands won't just be swinging an axe on the surface, as they would first need to solve the problem of being able to move around, so I'm responding to the analogy with humans who have to do the task without hands, but do have legs. For dolphins, we would need to start with the reference procedure where it's clear how dolphins with hands can do something.

To chop down a tree, you need to strike it repeatedly with an axe (this is what I assume you meant). To strike it repeatedly, you need to be able to strike it once. It's such actions as striking a tree with an axe once that I meant as something that I expect can be reduced.

Let's make the handle of the axe a much longer stick, and also attach another stick perpendicularly to control the tilt of the head of the axe, so that it's possible to make sure that the blade is turned in the correct direction without having to apply torque directly to the handle. The long handle can be placed on top of a third stick perpendicular to it, and ride along that third stick, with the end (knob) of the handle fixed in place. When it does so, the head of the axe swings. Now, if we let the head of the axe fall under its weight while guided by ("riding" on) the third stick, or alternatively pull it in order for the axe to gain the necessary speed, and use the second stick to direct the blade, the result is the axe head striking the tree with the blade at sufficient speed to dent it. Perhaps such method would be a hundred times slower, so that it would take a year to do a job that would otherwise take a day, and that is just what I meant by the process being much less efficient, more laborous.

Now imagine this creature as strictly waterbound

(Not sure what you mean by "strictly waterbound", though this distinction doesn't seem important for this discussion. The hypothetical considers creatures that are like dolphins in all relevant respects excepts they also have hands (maybe as additional retractable limbs, to preserve swimming capabilities). So they should be about as waterbound as dolphins. If this hypothetical allows technology, we could pose the more difficult problem of developing technology without the ability to surface even for a short time (which dolphins have).)

I think the basic problem here is that I have to prove a negative, which is, as we all know, impossible.

It's not impossible. Significant evidence of the negative will be obtained if performing a thorough investigation (which would be expected to solve the problem if it can be solved) fails to solve the problem. Applying this to flaws in particular steps, the useful goal is to show that something can't be done (that we won't find an alternative solution), not just that something won't work if done in a particular way.

For constructing a plan, I have another idea. Start with the simpler problem of developing technology as dolphins with hands. This hypothetical isolates the problem of dealing with underwater environment, from the problem of dealing with absence of hands.

Let's suppose that it's possible to solve this simpler problem. Then, I'm not sure that when we have a particular tiny operation that could be performed with hands (a step in the process of developing technology by dolphins with hands, such as smashing something with something else, or tying a knot), it's impossible to reproduce it without hands (much more laborously, slowly, using more people). Can you come up with a particular example of a very simple action that can be performed with hands (underwater, etc.), which doesn't look like it can be reduced to working without hands?

Mostly I expect creative surprises based on overall impression about the power of engineering. Let's try to do a bit of exploratory engineering, consider projects that include steps that are clearly suboptimal, but seem like they could do the trick. (A practicing engineer or ten years of planning would improve this dramatically, removing stupid assumptions and finding better alternatives; a hundred thousand years of actually working on the subprojects will do even better.)

Initially, power can be provided by pulling strong vines (some kind of seaweed will probably work) attached together. It should be possible to farm trees somewhere on the shoreline, if you don't mind waiting a few decades (not sure if there are any useful underwater plants, but there could be). A saw could be made of something like a shark jaw with vines attached to the sides, so that it can be dragged back and forth. This could be used to make wooden supporting structures that help with improving control of what kind of change is inflicted on the material by a saw. Eventually, incremental improvements in control and precision of saws would allow getting to something functionally similar to sawmills, bonecraft and woodcraft tools.

These enable screws, joints, jars and all kinds of basic mechanical components, which can be used in the construction of tools for controlling things on surfaces of rafts, so that in principle it becomes possible to do anything there given enough woodcrafting and bonecrafting work. At this point we also probably have fire and can use tides to power simple machinery, so that it's practical to create bigger controlled environments and study chemistry and materials. And we get concrete/cement to create watertight buildings and possibly canals with locks for land access. Something like ironsand or ores from surface exploration can be used to initially get metal and develop precision tools, at which point we get electricity and more powerful chemistry capable of extracting all kinds of things from available materials, however inefficiently. After that, there doesn't appear to be much difference from what's available to humans.

However, dolphins would hit a different filter, with their unfortunate body plan, lacking any type of fine manipulator limb whatsoever, making it infeasible to build complex tools.

Do you expect animals with human-like intelligence and dolphin-like bodies will fail to develop technological civilization? As a first approximation, I expect a community of modern human engineers (with basic technical background, but no specific knowledge) in dolphin bodies can manage to do that eventually, if they form a society conductive to long-term pursuit of the project. It's less clear if at human level this happens spontaneously, since it did take 200,000 years for humans with hands to get to technological civilization, and an additional difficulty could make it millions of years if intelligence is kept fixed.

(Assuming that machiavellian intelligence pressure can run further than it did with humans, machiavellian dolphins could at some point become even smarter than humans, which can be used to overcome the no-hands difficulty more effectively than human-level dolphins could. Alternatively, human-level dolphins can learn of selective breeding and create smarter dolphins irrespective of whether smarter dolphins would arise on their own.)