Please re-read. The numbers I gave are for fixed panels (i.e. where the tilt is fixed).

The only datacenter that needs heating is the one that's empty and thus either soon to be full or soon to file for bankruptcy. Also there's always quite a bit of unenclosed space between the panels and the roof, even in cases where the panels are flat rather than at the latitude tilt. So expecting the heat to transmit from the solar panel to the roof seems pretty unrealistic.

They are also not flat against the roof meaning there's a big layer of mobile air between them and the roof. So the heat does not transmit from the solar panel to the roof. Hence all that remains is the shading which does mean they considerably reduce the heat loading on the roof.

Actually it's 4.8 hours/day for a fixed tilt panel in the worst location of the 48 lower US states (St. Louis, Missouri). So almost any US location is guaranted an average of more than 5 sun-hours/day. Germany is indeed a bit worse but not as much as some think. Still Berlin can only count on about 3 sun-hours/day. That previous link also clearly shows that there's significant seasonal variation. This can be mitigated a bit by using the grid to connect to less affected places. Still a 100% renewable scenario does require the need for either lots of seasonal storage (the hardest kind), or lots of over-capacity.

So your counter example is a 100% renewable scenario that ran on 100% renewable for less than a day! And the next day that 100% renewable scenario ran on less than 100% renewable??? What this proves is that such a scenario cannot exist without energy storage.!

He's looking into what it takes to have a 100% renewable power grid. As mentioned in a previous post, providing storage for a day or more could well be needed in such a scenario. And if you read his analysis you'll see that whether you try to provide storage for a week or a mere day does not matter much: pumped hydro does not cut it in either case, by far.

You see to be unable to get past "we have a solution based on cheap natural gas that works now". But assuming natural gas will remain cheap, plentiful(*) and have no nasty environmental side effect is taking quite a risk.
(*)We have proven reserves for 56 years, at 2010 consumption levels, less if consumption increases at it historically does.

The feds don't give that $7500 subsidy because the brand of the car is 'Tesla'. They give it because it's an electric car. There's over a dozen models from BMW, Fiat, Ford, General Motors, Nissan and others that get the same subsidy. So spinning this subsidy as "the feds are helping Tesla" is a lie. Same thing for the state subsidies.

The Tesla Model S is a full-size luxury liftback which ranks second against other "Super Luxury" cars and has a 0-60 acceleration of 4.2 to 5.9 seconds (depending on model). The Hyundai Sonata is nowhere close to being in the same class seeing as it's a mid-size car and not as sporty with a 0-60 acceleration of 6.5 to 8.5 seconds. So again you're ill-informed or disingenuous.

Wow! I see neither the solar capacity assumption, the error or the attempt to mislead. I must be so dumb. Or did you mean there's no way PV can ever provide a significant portion of our electricity needs?

Given that it follows the section that started with "So that's dealt with the article that kicked off the discussion - now for the one you've linked", I'd say it is.

This relies on the assumption that fossil fuels will remain cheap and plentiful for decades to come, and that we don't need to reduce their use for environmental reasons. I think both are incorrect. It also assumes that building and operating spare gas turbines to deal with variable load is cheaper than using grid storage. That's correct for now but it's no reason not to look for better solutions. Cheap electricity storage would have other uses besides helping the grid deal with variable load and production. For instance it would help wind turbine and PV plants operators better monetize their production by not forcing them to sell on the spot market.

I'll let Americans speak for themselves.

Sure there's wind, but not necessarily plenty, and half the solar panels in the north of the US may be covered by snow. Which means that for a week you may have electricity production that's 25% below demand. So what do you do? Trust in the smart grid to automatically reduce demand by 25% for a full week? Forcefully reduce demand through brownouts? Or can we simply have two days of electricity storage to make up for it?

Now Tom Murphy's point is not to say that this is impossible, but that it's not trivial as some people (like the GP) seem to think. Hence if storage the route we want to take we have serious work and research ahead of us. Also just building the energy storage infrastructure (e.g. al that concrete for pumped hydro) will itself use up lots of energy, so we should do so while energy is still plentyful. His further point was that it would be much simpler to reduce demand by travelling less, eating less meat, better insulating buildings, heating them less, etc. While there's no doubt stuff to do to reduce demand he is definitely an outlier in how far he is willing to go to do so.

Of course a lot hinges on how much the spread of energy sources can limit the shortfall, and how long such shortfalls can last. My understanding is that neither of those are quite settled issues yet. Also that can be somewhat mitigated by overbuilding renewable production capacity, though if we find a good storage solution that might not be the most efficient use of resources. In any case researching grid storage makes sense to keep our options open, though personally I'm still rooting for the distributed car battery storage.

Tom Murphy is a physicist, not a journalist, and has obviously taken quite an interest on the subject of energy as can be seen from his blog. So he's certainly as knowelegeable as you or me as far as evaluating the rough feasability or cost of various solutions. And while I trust him to do the math right, I don't trust him about evaluating how far society can be changed. As for the Slashdot article, it was essentially content free: there's a thing called grid storage, batteries used for it have different requirements, some people are doing research on that topic. Is that news? Did anyone on Slashdot not know that already? Hard to get sucked in by a journalist that has nothing to say.

Pumped hydro works but just cannot be scaled to provide sufficient storage. Hence other solutions are needed. Actually it's likely nothing short of a combination of many approaches will be enough.

Wrong counter-example as it's not one I mentioned and it's irrelevant to current events.

Long discredited? That may be so but we still have lots of religious people who oppose teaching evolution or reproductive biology on a religious basis, disbelieve climate change in disproportionate numbers, believe the earth is about 6000 years old, or even, in some parts of the world, think that girls have no need for education.

Finally most religions require one to accept truths on faith, that is without objective reproducible proof. That's the anti-thesis of the scientific method.

If the block can be bypassed using UDP then the ISP made a 'big stupid error' as I mentioned. Their router should simply not forward any packet outside the local network until the customer provided his credentials. That covers IPv4 (TCP, UDP, ICMP, others), IPv6, and anything else, whether they support it or not. For ADSL it should be pretty easy to identify the customer's line and redirect anything coming from that line, leaving no possibility of escape. Customers who connect to their ISP through a shared medium, like cable or WiFi, there's an escape route which is to hack their hardware/software stack to impersonate another customer on that shared medium. But that's obviously illegal and furthermore there's no point for them to keep paying for Internet access in the first place.

So it's not a printer in any sense of the word. Great start for that article. The rest really goes downhill from there. Shouldn't it have been published on the 1st of April?