Check out the SNAP-19.

I know of it - it powered the Pioneer spacecraft in the early 1970s. Can I get one today? No. Can anyone take the drawings for it (assuming they can be found) and manufacture new ones for signficantly less than the cost of creating a modern design? Probably not. Could the ESA do it? Probably not.

I reiterate the points in my earlier post. Please read further.

How long it lasts is dictated by nuclear physics (the half life of Pu-238.)

The decay in power output is only partly due to the half-life of the nuclear fuel. An effect of roughly equal magnitude is the degradation of the thermocouple junctions themselves. For instance, each Voyager spacecraft's RTG had about 470 W electrical output at launch in the 1970s. As of 2008, about one half of a half-life later, the electrical output had declined to about 285 W. I'm sure that the thermal output of the RTGs is following the half life of Pu-238 just as one would expect, but the useful life is limited by other considerations.

But what would happen if it would be heated up and worn down in a low angle orbital reentry? It could be subjected to melting/burning temperatures for many minutes

The Lunar Module from Apollo 13 carried an RTG, which was intended to power surface instruments on the Moon. It re-entered, along with the rest of Aquarius at around 25,000 mph, at roughly the same entry angle as the Odysee command module, which is intentionally shallow to bleed off maximum velocity with survivable g-forces.

That Pu-238 cask is now sitting at the bottom of the Pacific ocean. All seems well.

Though, this is its thermal output. If you consider that Seebeck generators have a 10% efficiency, you could get 32W electrical out of 640g of PU-238. Let's account for the 10 years trip, so let's make it 1kg

The weight of the actual plutonium isotope is but a small fraction of the weight of the finished RTG. There's the thermocouple wires, the iridium cladding, the graphite casing, the metallic casing, etc. No one has made an RTG with just 10s of watts of output since the 1970s, but those designs weighed in at 10-20 kg, which is about the same mass as Philae's (solar+battery) power system.

Having an RTG would not obviate the need for actual batteries. The RTG is good for providing average power, but the peak power draw is much higher than that, and batteries are used for that purpose.

Plutonium 238 is really not that dangerous. It is an alpha emitter, and carried in a water insoluble form.

Radiation aside, it is also an extremely toxic heavy metal.

That said, it isn't safety concerns that kept RTGs of Rosetta/Philae. If the project had determined that the best way to achieve the science and program goals was through an RTG, they would have launched with an RTG. The safety concerns about exploding rockets or re-entering spacecraft didn't keep Curiosity, Cassini, or New Horizons from launching. Rather, if ever the engineering design and program management discussed using an RTG, they decided against it for much more pragmatic reasons (engineering, cost/benefit analysis, program risk, etc.)

A 32W RTG would generate about 600W of waste heat, something that is easy to radiate over 3m2 into space, assuming reasonable operating temperatures for the probe (and actually, a smaller RTG is sufficient)

I think this is something that isn't particularly well appreciated - a 32-W RTG does not exist. All the designs for recent spacecraft have been on the order of 100-300 W (electric).

I practically burst out laughing when the article gets around to introducing the notion of powering Philae using an RTG. The image that the author dramatically inserts at this point is the RTG for the Curiosity rover - an assembly that is itself about the size of the entire Philae spacecraft! Cassini, Ulysses, New Horizons - all of the recent RTG-powered probes used the same design, one that is entirely the wrong size for Philae.

For Philae to use an RTG, it would need to have been a new design (something the ESA has no experience in) - a development that could have cost more than the Philae lander itself. Even with a new design, they would have needed to secure the Pu-238 to power it (assuming they didn't use some other isotope, which would have been yet another costly design effort). When the craft was being designed and built, the supply of Pu-238 was already more or less spoken for, and it would have been an enormous program risk for them to commit to an RTG without a guarantee that they could fill it.

Hearing about lasers and raindrops puts me in mind of this recent XKCD What-If posting.

how is this ion engine more efficient than just squirting a small amount of pressurized gas out of the tank instead

It has to do with how quickly you can throw the propellant - how much momentum you can impart to it, which in return imparts a certain change in momentum to the rest of the satellite (delta-v). With conventional satellite propulsion, like fuel+oxidizer rockets or monopropellant thrusters, the energy available to impart that momentum is chemically based. That is, the propellants undergo a chemical reaction, get hot and/or change phase into a gas, and nozzles force that gas to exit at some velocity. Details vary with engine and nozzle design, but there are limits on how much thrust you can get each fuel type. Mass in, reaction energy, mass*velocity (momentum) out. Rocket designers measure this "efficiency" with a quantity called specific impulse (measured in units of seconds) For a given mass of fuel, you can pretty quickly calculate what the total delta-v the satellite has available to it.

Ion engines can impart much higher velocities to the "fuel" than chemical rockets, in part because they are using electrical energy (of which there is an arbitrarily large supply) rather than whatever you can get from chemical reactions. Again, the details vary based on the design, but ion engines tend to have specific impulses much higher than chemical rockets. The actual thrust (i.e., total force) from an ion engine tends to be miniscule, but is provided very efficiently, and can be produced for days or weeks at a time.

Google has been working on that, it's called Flu Trends. But it hasn't really proven itself out yet. See my post below.

Google tried (is still trying?) to track the spread of influenza, by watching the trends in searches for information about the disease. It's a very interesting bit of work, but as I recall, failed to be meaningfully predictive. The trouble is, there are lots of prosaic reasons why someone might search out information about the flu (or any other disease) other than actually having it. Separating that noise (general interest in the flu) from the genuine signal (particular interest from people who are infected). Doesn't mean it can't work, just that it hasn't been made to work yet.

well played.

proper Stirling engines or steam turbines are not popular in space for some reason

R&D on a nuclear-powered stirling engine for space is ongoing. It's not they they aren't popular, per se, its just that they are a very difficult engineering problem. How many devices with continuously moving parts do you know that operate maintenance-free for years or decades? It's not impossible, but is really hard.

As others have mentioned: RTGs are difficult to come by ($$$, rationed resource) compared to solar panels. Generally, they only get used when there is no other way to power the spacecraft to meet the science objectives.

The present RTG designs for spacecraft are all 1-2 orders of magnitude larger than what Philae would need.

One other thing I would note is that RTGs have useful lifespans measured in years to decades. Philae hasn't been designed to operate for that long (its long sleep until it arrived notwithstanding). What is more, even if it has electrical power (solar, nuke, or otherwise) to operate indefinitely, it will be cooked by the Sun around the time of the comet's closest approach. So an RTG would vastly outlive the useful lifespan on Philae. Given that RTG material is scarce, this would be a wasteful use.

reminds me of a similar situation that came up in an introductory CS course at Dartmouth years ago. A bunch of students, possibly abetted by the TAs, ended up with the exact same code for some problem or other. The professor (a visiting professor, if that makes any difference) went ballistic and reported every student, intent on getting all of them suspended or expelled.

In the end, the charges were dropped. It would have been a very difficult situation to adjudicate - the behavior of the students, TAs, and professor was bad all around, and some might say that it all cancelled out. Some definitely were cheating (i.e., malicious intent), and others just got caught up in stupid behavior (i.e., ignorant), and it would have been difficult to separate the two.

The usage of "United States of America" as either singular or plural has shifted over the years. The guidelines you provide are good for the general case of collected objects, but "USA" seems to be a particular case that (sometimes, maybe, mostly) breaks the rule.