And, of course, there's the "separation of powers" thing. The TSA is under the executive branch, so they can't really do anything to judges - the judicial branch has to police itself. For pretty much the same reason that the executive branch can't do anything to Congress - the police are constitutionally not allowed to arrest a congressman, because if they could, then the executive branch could (legally) take over the government by physically arresting Congress to prevent votes, etc.

You can make a better "gun" than the Liberator using a piece of wood and a drill, faster and cheaper than a 3D print. The US doesn't suffer from a shortage of guns or the ability to make guns.

The only reason to 3D print a gun is because you really like guns and want to use 3D printing to get some press.

Sure, there would be regular supply schedules. But that means that the part arrives weeks or months after it's ordered. Which means either maintaining a large inventory of spares before they're needed, or waiting weeks or months while equipment is out of commission.

With an SLS printer, they could have the part in a few hours. So if the value of time is high, it's worth the cost of the SLS part.

You're right that SLS printers are slow and expensive, but they're still faster than physically shipping parts (hours vs. days), and if the value of time is high then it's worth using an expensive process to get the part faster. And it lets the unit be more self-sufficient, which is valuable when traditional shipping breaks down - many wars have been lost over control over supply lines.

And an SLS printer and metal powder is much simpler to warehouse and keep supplied than a complete inventory of all needed metal parts. I suspect that the complete supply chain optimization is pretty impressive when you add it all up. Mass production is very efficient, but add in massive manufacturer markup, then all of the wearhousing, shipping, security, inventory, etc., logistics to manage and move everything around, and "print in the field on demand" is probably pretty appealing.

I'd bet that the the only thing holding it up would be negotiations with manufacturers - I'm sure that the manufacturers will want to get paid tons of money even if all they're doing is licensing the design files (i.e. not actually making or shipping anything). So the $500 specialized wrench would continue to cost $500.

But if the army guys had SLS printers in the field, how rapidly do you think they'd model their own parts that are better than the ones from the vendor? Those guys are quite ingenious at making things work with whatever they have available, and with an SLS printer, they sky's the limit!

There are lots of people doing casting from 3D prints. You can 3D print the master in PLA, then make a plaster cast, burn out the PLA, and pour in metal. For example, http://3dtopo.com/lostPLA/ . It works, it's just more dangerous and complex than most people want to deal with.

Not just 'tried to', but actually delivered. Thinking Machines CM-1 and CM-2 had routers on chip with 32 CPUs per chip. In the hybercube architecture, the 5 lowest bits of CPU address routed on-chip, and the rest of the CPU address routed between chips. It worked quite well, and was the fastest computer on the planet for several years running.

+1, Interesting.

(Thanks!)

I guess "fully steerable" means "within ~20 of zenith" to you. But that's not what the rest of the world considers "fully steerable".

Oops. That was in reference to his Green Bank Telescope link - Arecibo being only partially steerable...

The Extremely Large Telescope is a compromise - what they really wanted was the 100m-diameter Overwhelmingly Large Telescope.

Their naming committees are either entirely humourless or gloriously taking the piss.

The worlds largest single dish telescope is still the Green Bank Telescope (http://en.wikipedia.org/wiki/Green_Bank_Telescope), which at 100m is ~6x the size.

World's largest fully-steerable single-dish telescope - the Arecibo Observatory is larger still at a diameter of 300m! (Impressive Arecibo exploration video here. The thing's sodding enormous.)

I went looking for the largest diameter multi-dish radio telescope. It looks like the biggest terrestrial 'telescope' is the Global VLBI system created by combining the European VLBI Network with the US Very Long Baseline Array - it's like some massive team of superheroes combining to save the Earth from some terrible secret of space. Or whatever. Apparently they can also add space-based telescopes when that just isn't enough. Which, quite frankly, is showing off...

My thoughts when seeing one of the beautiful, 10m diameter Keck optical telescopes up close a few years ago? I've had full control of a telescope bigger than that.

Radio Astronomers: Compensating For Something.

Sounds like pretty much every time there's a new industry standard, where the major players all come up with their own incompatible option, trying to be the one that wins and gets to charge everyone else licensing fees for their patents and trademarks. And so, as usual, the innovative new field is fragmented, confusing consumers, wasting money, and delaying or even killing the new industry. These sorts of format wars happen so often that I can only think of one case (CDs) where it didn't happen. You think that after wasting so many years, and $billions, on these pissing contests, and seeing that the one time they didn't screw it up it was a huge success making everyone rich for decades, that businesses would learn that it's better to cooperate on standards rather than compete. But I guess they're so competitive that they do it every when it's a consistently bad strategy.

You can learn some languages that have clean/consistent designs in a few hours. Lisp, C, Lua, Smalltalk ... those languages are simple, clean languages that you can learn the syntax of in a few hours.

That begin said, to really get work done you need to understand the runtime environment, GUI and other frameworks, database/persistence, messaging, performance characteristics, etc., and that's vastly more complex than the syntax of the language. Even if you learn Python in a few hours (you can), there's no way you're going to understand the frameworks and how to get real work done in a real application in a few hours. Even in an environment that takes care of most of the details (like Google App Engine) it takes a few weeks to wrap your head around the High Availability Data Store, Google's APIs, etc.

I suppose you have a chance, if all your code needs to do is run on a command line, process stdin and generate stdout with no persistence. But that doesn't describe most applications these days.

You're picking random numbers to try to make a comparison that's not meaningful. 292 groups applying for tax-exempt status had names contained "tea party", "patriot", or "9/12", who were all given more scrutiny, and only 20 groups applying for tax-exempt status had names contained "progress" or "progressive". Of course, those numbers you care about are only about a third of the groups given deeper scrutiny - the large majority of groups investigated for possibly being political groups (and this not allowed to claim tax exempt status and hide their donor lists) weren't right-wing groups, they had terms like "Democrat", "Occupy" or "Israel" in their names, and only a third of the groups were right-wing groups, so the scrutiny wasn't politically biased against right-wing groups. If anything, it affected more left-wing groups than right. It was still a bad idea to use a list of terms to trigger deeper investigation, of course, but as the "BOLO list" system was put in place by a conservative Republican who was trying to make the system more efficient and consistent, and was applied to groups across the political spectrum, and the outcome was that only a third of the groups affected were right-wing, the evidence suggest that it wasn't politically biased against the right-wing.

The political bias that is most obvious is that the Inspector General's office in charge of the IRS audit had been asked by House Oversight Chairman Darrell Issa (R-Calif.) "to narrowly focus on Tea Party organizations". And, of course, Issa kept most of the testimony secret, selecting a few bits to try to distort the program as anti-right-wing.

The reality is that using a list of terms to watch for is a bad mechanism, because it's substituting a mechanical rule for human judgement in making a determination about a group's being political vs. social, and that's wrong. But if you ignore the rhetoric, the facts don't support the accusation that the BOLO lists were aimed purely at right-wing groups. The real problem is that Congress passed a law requiring the IRS to make an extremely vague determination ("primarily political") so the IRS came up with a system for making that determination that pissed people off.

Looking at the lists, I can't see how most of them were ever granted tax exempt status. How can a group with a political party in its name, that raises money for and donates time to political candidates, not be a political group? The IRS should have rejected far more of the groups, both Republicans/Tea Party and Democrats, and instead of seems like all they did was ask a lot of questions and then approve almost every application, which seems a bit pathetic.

+1 this. The only time I've ever had Mac filesystem problems was when there were unexpected power loss. When you lose power while writing to the drive, bad things can happen. But I've not seen even that since Mac OS 7 or so. :-)

I used to work in supercomputing, and with terabytes of RAM and Petabytes of data I/O, you *bed* everything had ECC and parity bits every step of the way (yeah, extra parity bits in the RAM). And cosmic rays really do flip bits in RAM from time to time, and when you're running a $10M machine running a 2 month computation, you really do care about being able to detect the error, restore the machine's state to the previous snapshot, and keep running.

It's amusing to me that consumers now have enough data that this stuff starts affecting them. It'll be interesting to see if consumers start paying extra for the reliability.