Southwest. Above Jake's Seafood (which, incidentally, serves terrible food at absurd prices for the benefit of the ambiance of an old brick building).

Portland is cool, yes. But that's mostly down to the bookshops and tea shops. Temperature-wise, it doesn't get "hot" per-se, but it does get very humid. And the air is horribly polluted. Spent the time moving up there reading about dangerously high levels of mercury in the air, the deadly pollutants in the river, the partially dismantled nuclear reactor and highly toxic soil (the reactor has since gone, the soil remains drenched in contaminants), the extremely high levels of acid rain due to excessive cars (which are driven by maniacs) and the lethal toxins flowing through the rivers that have been built over to level out the ground.

In short, I landed there a nervous wreck.

Things didn't improve. I saw more dead bodies (yes, dead bodies) in Portland and heard more gunfire in my five years there than I heard in the suburbs of Manchester, England, in 27 years. You will find, if the archives let you get back that far, that I was almost normal before that time.

What about a TARDIS?

1955. The Manchester Computing Centre was designed to be one gigantic heat sink for their computers in the basement, using simple convection currents, ultra-large corridors and strategically-placed doors to regulate the temperature. It worked ok. Not great, but well enough. The computers generated enormous heat all year round, reducing the need for heating in winter. (Manchester winters can be bitingly cold, as the Romans discovered. Less so, now that Global Warming has screwed the weather systems up.)

The design that Oregon is using is several steps up, yes, but is basically designed on the same principles and uses essentially the same set of tools to achieve the results. Nobody quite knows the thermal properties of the location Alan Turing built the Manchester Baby in, the laboratory was demolished a long time ago. Bastards. However, we know where his successors worked, because that's the location of the MCC/NCC. A very unpleasant building, ugly as hell, but "functional" for the purpose for which it was designed. Nobody is saying the building never got hot - it did - but the computers didn't generally burst into flames, which they would have done if there had been no cooling at all.

Good point. It could be possible to tamper with the embassy's water supply in such a way that the effects would only be felt by someone who *lives* there...

I might get to see the first car get diverted into a schoolbus to avoid a 50-million-dollar superduperhypercar. I'll have to dress for the occasion with my best fingerless gloves and head-worn goggles.

First you'd need to root the car and run "echo 1 > /dev/morality/evil"

Well, there are tons of trolls on SA. They just spend their time trolling other sites.

Some of the best humor in the world is a result of idiots on Facebook not recognizing satire!

I wish I could find one particular Facebook screenshot of a bunch of right-wingers who took the "abortionplex" article seriously. Just pure gold.

And Europeans know very well about slavery and exploitation. They were the masters of it at the time.

Let's start with basics. Message-passing is not master-slave because it can be instigated in any direction. If you look at PIC Express 2.1, you see a very clear design - nodes at the top are masters, nodes at the bottom are slaves, masters cannot talk to masters, slaves cannot talk with slaves, only devices with bus master support can be masters. Very simple, totally useless.

Ok, what specifically do I mean by message passing? I mean, very specifically, a non-blocking, asynchronous routable protocol that contains an operation and a data block as an operand (think: microkernels, MPI-3). If you're clever, the operand is self-describing (think: CDF) because that lets you have overloaded functions.

The CPU is a bit naff, really. I mean, at least some operations can be pushed into a Processor In Memory, you have a fancy coprocessor for maths that you're repeatedly (and expensively) calling to create the functions that exist as a limited subset in FFTW, BLAS and LAPack. Put all three, in optimized form, along with your basic maths operations into a larger piece of silicon. Voila, massive speed boost.

But now let's totally eliminate the barrier between graphics, sound and all other processors. Instead of limited communications channels and local memory, have distributed shared memory (DSM) and totally free communication between everything. Thus, memory can open a connection to the GPU, the GPU can talk to the disk, Ethernet cards can write direct to buffers rather than going via software (RDMA and OpenSockets concepts, just generalized).

You now have a totally open network, closer to Ethernet than PCI or HyperTransport in architecture, but closer to C++ or Java in protocol, since the data type determines the operation.

What room, in such a design, for a CPU? Everything can be outsourced.

Now, move onto Wafer Scale Integration. We can certainly build single wafers that can take this entire design. Memory and compute elements, instead of segregated, are mixed. Add some pipelining and you have an arrangement that could blow most computer designs out the water.

Extrapolate this further. Instead of large chunks of silicon talking to each other, since the protocol is entirely routable, get as close to individual compute elements as you can. Have the router elements take care of heat and congestion issues, rather than compilers. Since packet headers can contain whatever label information you want, you have a notion of processes with independent storage.

It doesn't (or shouldn't) take long to figure out that a true network, rather than a bus, architecture will let you move chunks
of the operating system (which is just a virtual machine, anyway) into the physical computer, eliminating the need for running an expensive bit of simulation.

And this is marketspeak? Marketspeak for what? Name me a market that wants to eliminate complexity and abandon planned obsolescence in favour of a schizophrenic cross between a parallel Turing machine, a vector computer and a Beowulf cluster.

Don't forget the mercury output as well. It's largely thanks to coal that a few pounds of swordfish steaks now contain (on average) as much mercury as a typical CFL bulb, and the level of mercury in our oceans continues to slowly increase over time.

OpenCL is highly specific in application. Likewise, RDMA and Ethernet Offloading are highly specific for networking, SCSI is highly specific for disks, and so on.

But it's all utterly absurd. As soon as you stop thinking in terms of hierarchies and start thinking in terms of heterogeneous networks of specialized nodes, you soon realize that each node probably wants a highly specialized environment tailored to what it does best, but that for the rest, it's just message passing. You don't need masters, you don't need slaves. You need bus switches with a bit more oomph (they'd need to be bidirectional, support windowing and handle multipath routing where shortest route may be congested).

Above all, you need message passing that is wholly target-independent since you've no friggin' clue what the target will actually be in a heterogeneous environment.

Hemp turns out to make a superb battery. Far better than graphene and Li-Ion. I see no problem with developing batteries capable of supporting sub-zero computing needs.

Besides, why shouldn't public transport support mains? There's plenty of space outside for solar panels, plenty of interior room to tap off power from the engine. It's very... antiquarian... to assume something the size of a bus or train couldn't handle 240V at 13 amps (the levels required in civilized countries).

Very true, but without it, we're doomed to reinventing wheels, redoing research and coming up with suboptimal solutions that are harder to program, harder to maintain and bloated with helper functions that would have come as standard otherwise.

Such a table can be written once then updated every 5 years. Reading it simply amounts to feeding into a parametric search routine what you know you will need to be able to do. You will then get a shortlist of languages ideal for the task.

Now, it comes down to two simple questions: are your requirements ever stable enough or clear enough for such a shortlist to be useful? Do you risk overoptimizing on a set of criteria that may have no resemblance to the reality of the problem or the reality of any solution the customer will sign off on?

If the answers are "yes"and "no" respectively, I'll start the list today.