When I mentioned breakout boards, I meant pre-built ones. With the chip already soldered on it.

With regards to hot air rework stations... I said heat gun. You know, the kind you use to strip paint? A heat gun, some tinfoil, and a decent PCB, and you too can solder SMD. Making your own PCBs is really cheap Like I said, about $10 for 10 if you can keep them to 5cmx5cm. See http://imall.iteadstudio.com/open-pcb/pcb-prototyping.html for more info.

You still haven't demonstrated what kind of abuse breaks an arm, and from which manufacturer. All you've done is demonstrated what kind of abuse an AVR can take. That has absolutely no bearing on what an ARM can handle.

So what you're saying is that AVRs are for people who aren't good at electronics? That's fine. The next step up for them after they learn how to connect an LED can be to get an ARM board.

Surface mount is a moot point in the face of an inexpensive breakout board, unless you're looking at a size-constrained application. You can have your own PCBs manufactured professionally for $10 if they're small (5x5cm), so SMD parts are viable as long as the pitch isn't too small--I've soldered small SMD parts many times with both a heat-gun and a soldering iron. I like the heat-gun better, but the soldering iron is more commonly available. That said, if Arduino is a contender, then use of breakout boards is a non-issue.

That's a pretty sweeping statement. Do you have any evidence to back that up? You know that NXP's line of ARM micros are all 5V tolerant, right? And ST's ARM lineup all have at least *some* 5V tolerant pins, most of them are mostly 5V tolerant. The STM32F4 which is on the discovery board has 138 of 140 5V tolerant pins. TI's micro that's on the launchpad also has all 5V tolerant I/O.

5V tolerance is a non-issue.

If this is an issue, you're doing it wrong. VCC--|>|---/\/\/\---MCU pin. With 5V tolerant I/O, you no longer have a problem.

See above comment about breakouts.

Just because you don't need it for a particular application doesn't mean that having it available is bad. Maybe someone *does* need that. Then they have it available.

There's nothing wrong, per se, with AVRs or PICs. It's just that the price/performance tradeoff isn't very good in the face of other options.

I still don't understand why people are focused on PICs and AVRs. ARM has had better functionality and pricing (starting at the mid-range; low range is still dominated by 8-bit) and better peripherals for at least 5 years now.

TI Stellaris launchpad: $5, 80MHz, 32-bit ARM Cortex-M4 CPU with floating point, 256Kbytes of 100,000 write-erase cycle FLASH and many peripherals such as 1MSPS ADCs, eight UARTs, four SPIs, four I2Cs, USB & up to 27 timers, some configurable up to 64-bits. Integrated in-circuit debugger.

Each of the following also have an integrated in-circuit debugger which is compatible with OpenOCD
STM32F0DISCOVERY: $8 48MHz ARM Cortex-M0, 64 Kbytes of flash and 8 Kbytes of SRAM, standard communication interfaces (up to two I2Cs, two SPIs, one I2S, one HDMI CEC, and up to two USARTs), one 12-bit ADC, one 12-bit DAC, up to five general-purpose 16-bit timers, a 32-bit timer and an advanced-control PWM timer.
STM32VLDISCOVERY: $9.90 24MHz ARM Cortex-M3, 128KB Flash, 8KB SRAM, standard communication interfaces (up to two I2Cs, two SPIs, one HDMI CEC, and up to three USARTs), one 12-bit ADC, two 12-bit DACs, up to six general-purpose 16-bit timers and an advanced-control PWM timer.
STM32F3DISCOVERY: $10.90 72MHz ARM Cortex-M4, with FPU, 256KB Flash, 48KB SRAM, up to four fast 12-bit ADCs (5 Msps), up to seven comparators, up to four operational amplifiers, up to two DAC channels, a low-power RTC, up to five general-purpose 16-bit timers, one general-purpose 32-bit timer, and two timers dedicated to motor control. They also feature standard and advanced communication interfaces: up to two I2Cs, up to three SPIs (two SPIs are with multiplexed full-duplex I2Ss on STM32F303xB/STM32F303xC devices), three USARTs, up to two UARTs, CAN and USB. To achieve audio class accuracy, the I2S peripherals can be clocked via an external PLL.
STM32F4DISCOVERY: $14.90 168MHz ARM Cortex-M4 with FPU, 1MB Flash, 192KB SRAM, and way too many peripherals to list here.

All the above are supported by GCC and OpenOCD.
With prices, capability, and development tool support like that, why would you use an 8-bit micro? It doesn't give you the same support that Arduino does. You have to learn how the peripherals work and you have to write your own interfaces to things like ADCs, but the examples are pretty complete.

You're quite right. Trademarks are both geographically restricted and market restricted. If you're in a different market or a different location, the trademark doesn't apply.

And you have to register everywhere that you do business for the trademark to be valid in that region.

Trademark doesn't work like that. Trademarks are only valid in particular regions. If Gaymer's is a UK cider producer since 1770AD, then they most likely hold a trademark in the UK. That doesn't mean that another entity can't register the same trademark somewhere else.

Only in strings. We migrated to Words in the late '80s, then DWords in the '90s.

You can't. Android has a re-implementation of libc, which is missing some things you'd expect. Like any of the normal IPC mechanisms. If you want to port GNU to Andoird, you have to bring your own libc with you.

It would seem to me, then, that the charge should be fraud, not antitrust.

Back-slash-dot? Is that where you don't bother reading TFA? I read before I comment, anyway.

It actually continues to work for you even after you leave. Adding renewable energy generation and high efficiency heating/cooling (geothermal) to your home increases your property value, which gives you the option to do the same again, or buy one with the work already done.

It looks like I can buy solar modules for a minimum cost of $1/Watt.

Assume an energy cost of $0.1/kWh. Assume an average of 12 hours of sunlight per day and a 50% of maximum average intensity.
$0.1/kWh * 1 year / 12 * 50% * 12 hours/24 hours = $0.01826

The monthly value that a solar cell generates is $0.01826/watt month.

Assume a yearly interest rate of 5% (monthly is 0.4074%)

Since the cost of a solar cell is $1/watt, work out the number of months that a 1W solar cell must run for to generate $1.
PV = A/i (1-1/(1+i)^n)
PV = $1, A = $0.01826, i = 0.004074

n = 62 months = 5.17 years

The warranty on the reference cell is 10 years product workmanship, 25 years linear power.

So the value of the cell over its 25-year life span is $3.15/watt, with a cost of $1/watt.

This all neglects installation and grid-tie costs, but 50% average illumination per daylight-hour is conservative in most areas. Solar cells ARE worthwhile TODAY and WITHOUT government subsidies.

Efficiencies in solar cells are irrelevant. The only thing that matters is the $/Watt.
Reference Solar Cell: http://www.affordable-solar.com/store/solar-panels/CSI-CS6P-245P-245W-Solar-Panel-STD-Frame

And I was thinking of a mechanism to let me play Rick Astley clips without anyone being able to fight back.

While I'm with you on the question of performance, I'd also question the suitability of FPGAs, both as an "open source" platform and as a learning tool for anything below university level courses. FPGAs are about as closed as it gets when it comes to hardware platforms. The verilog/VHDL compilers are, generally, closed source. I know there's an open one or two, but Mentor Graphics, Xilinx, and Altera all ship closed source compilers. The place and route algorithms that are used are all patented and closed source. The architecture of the FPGA itself is patented and closed source.

So, what, exactly, is the point of using an "open" processor on an FPGA? To make everything harder to do?

If you're really looking for a Free/Open processor, then your best bet is to put your money where your mouth is and back opencores.org in producing an ASIC version of the OpenRISC 1000. Even then, it's still built on a proprietary process in a fab, where you can't even get the technology files required to layout the processor without signing an NDA.

Here's the sad truth of it. You're dealing with a proprietary process anywhere from the chip level down. You simply cannot complain about not having open silicon and be taken seriously. Here's how it works:

If you want to make a chip, the first thing you have to do is find a design. Now you can make your own, and open source it, or you can get a pre-made design. If you choose to use an open-source design, then you're good--so far--but you'll have a significant performance lag behind the proprietary options. This goes double for video processing, memory controllers, buses, etc., etc.

Next, you need to find a fab who will make the chips for you. Here's where it gets bad. Even 180nm fabs consider their processes to be trade secrets, so that you have to sign an NDA just to get a process description file from the fab--this means your layout is, perforce, closed source.

Even if you somehow find a fab which will allow you to open the technology file, the placement and routing software for VLSI design is all closed source and patented. This is because place & route is a HARD problem. NP Hard, in fact.

So what it comes down to is this: until the homecmos people get their process going, you're stuck with something proprietary at some level. So then how much proprietary stuff is tolerable?

The Raspberry Pi Foundation had the goal of being bringing computing in a low cost package for education. The tradeoffs required to use open designs for the processor are quite steep: e.g. it would be a colossal time investment to get Linux running on a non-standard--read: non-proprietary--SoC. Using some proprietary chips to get there seems reasonable, so long as the OS doesn't become proprietary. The GPU blob is unfortunate, but not unexpected, particularly if you want decent performance.

How does your laptop computer calculate remaining battery life?

Two possibilities: 1) it has a voltage measurement, 2) it has a "fuel gauge" chip. For 1), It measures the voltage on the battery and compares that to a calibrated charge remaining vs. voltage curve. For 2), the chip reports the remaining capacity. Then divide by a sliding frame average of current consumption.

How does your browser calculate remaining download time?

It divides the remaining file size by a sliding frame average of the download speed.

How does your tablet distinguish between gestures?

It records all the points of a given "touch" (the point of a touch is the vector average of the whole touch zone. Think of it as the centre of gravity). It then does linear regression on the points, giving it a line which matches the average. It compares the linear regression and the actual data to find an error level, and divides the line segment up into smaller segments until they are all within some acceptable error level. It then passes this information off to a topological processing system, which could be as simple as a set of table lookups, but more likely is just a bunch of vector arithmetic.

The point here is not whether any of these examples is accurate or not, it's that each of them is plausible. You can do an enormous amount in programming without calculus.

But is it necessary? I'd argue it's field and application dependent. It's remarkably necessary for numerical methods (both for error and complexity); but those only matter if you're writing your own algorithms and you can get a long way without understanding the error in a system. It's necessary for simulation software, and there are more examples. Anywhere that a system is critical, and connected to the physical world, I'd want them to know the calculus required to accurately model their interface.

Another point, is that Computer Algebra Systems are faster and more accurate at most calculus than you will ever be; however they still require you to set up the problem so that they can do the algebra, integration, and differentiation for you.

The internet will be largely unaffected: the amorphous network of routers which makes up the internet is specifically designed to route around damage like this. The internet will be fine. The web, on the other hand, could suffer a large loss of access to content.