Dan's Data letters #203Publication date: 11-Nov-2008.
Last modified 03-Dec-2011.
I'm not too sure if you do articles on anything political (or even if you are aware of this), however Senator Stephen Conroy plans to put in legislation to filter the [Australian] Web:
What is your take on the effects of this proposal?
If implemented, will it work as planned? Will it break the Internet?
The "clean feed" idea aims to protect the public from the unavoidable scourge of child pornography that bedevils every Web user. I'm sure everyone reading this has experienced it - there you are, peacefully reading Cute Overload or something, when all of a sudden you're assailed by thousands of pop-up images of children being abused. Happens all the time. Obviously, mandatory filtering of every Australian Internet connection is the only way to stop it.
The clean feed also aims to protect Australians from "illegal" Web sites. Some might foolishly suggest that whether a site is legal or not ought to be a matter for the courts to decide, not for a secret blacklist made by unaccountable people who add sites to it in undisclosed ways. Such complaints, of course, are very probably coming from people who're in the kiddy-porn business. It stands to reason.
Like every other compulsory-Net-filter idea, this one enjoys wide bipartisan support among people who haven't the slightest idea what they're talking about. It seems to be a sure-fire vote winner, too, because people are always very happy to vote for something that's alleged to in some vague way stop kiddy porn, even if almost none of them actually want to filter their own Internet access.
(Won't they be startled if the blocker actually gets created, and then they can't read Gizmodo any more because someone once mentioned a modchip there?)
Before that can happen, though, there's some sort of pilot program, which will of course be a dismal failure, just like the pre-pilot-program test was. Because no, this system can't work as planned. Actual Internet criminals will just start using encryption (the sensible ones already are, of course).
The only way to stop that would be to block encrypted sessions of all sorts, whereupon nobody will be able to use online stores or Internet banking any more, and you can probably kiss your VPNs and Skype goodbye too. Which would indeed pretty much qualify as "breaking the Internet", if you ask me.
In the unlikely event that this lurching horror of an idea actually gets going, I think the best way to attack it would be with a class-action lawsuit from Australians with legitimate Web sites that're blocked, as many doubtless will be.
The blacklist, you see, implies that anybody on it is doing something illegal. Most probably, to listen to the promoters of the scheme, making or distributing child pornography.
When people making Linux ISOs available for download, or writing scholarly papers about "Lolita", or doing something that does not even to this extent resemble "illegal activity", get blacklisted, I hope they sue for defamation.
I wrote some more about the Clean Feed debacle-in-progress here.
(See also my ancient review of Pornsweeper. It doesn't look as if the ISP-level filters work much better than Pornsweeper did eight years ago. This doesn't surprise me; not only is effective content filtering a very hard problem indeed, but there isn't actually much selective pressure to make filter products any better. That's because it still doesn't matter whether censorware works or not.)
As you have a handle on most things and an armoury of tools to match, do you have any suggestions for easily removing screw-in bulbs from their holders?
I'm talking specifically about the 60W 60mm spots with the E27 fitting in those recessed housings where you cant get your fingers on the sides. I've replaced three this month and each one was fun-and-games (one actually shattered in my fingers).
Should I try putting some type of grease on the threads when I insert them (though I'm worried about fire/odours), or is there a tool for this situation?
I don't have a lot of experience with screw-in light bulbs, because bulbs here in Australia usually have a bayonet fitting, which pretty much can't seize up, even if badly corroded.
If I had to do it, though, my first tactic would be one big or a couple of small suction cups. Just stick 'em on the face of the bulb, and you ought to be able to get enough traction on the bulb to unscrew it. Suction cups about an inch across with a metal hook for hanging things can be had for a couple of bucks for a dozen or more from discount stores.
(See also "Removal of 100-Watt Electric Bulb from Rectum", a classic of medical improvisation.)
If a bulb actually breaks, the Standard Removal Tool is a raw potato. Just stab it onto the broken parts - maybe squirt a bit of penetrating oil into the threads first - and there's a good chance you'll be able to get the thing out. You can also use two pairs of pliers to grab opposite sides of the stump, but only at the cost of breaking off even more glass.
(After this page went up, a reader pointed out that you can get purpose-built light bulb changers, gripper doodads that you can hold in your hand or attach to the end of a broomstick. And yes, you can get ones with narrow fingers or suction pads that will work on tight-fitting floodlight bulbs. Here's one example on Amazon, here and here are kits that include a pole, and here and here are specific flood-bulb changers.)
And yes, lubricating the threads when you screw in a fresh bulb can help, too.
I think the major problem with using ordinary mineral or silicone oil for this purpose wouldn't be fire - one drop of oil isn't much of a fire risk - but the probability that the oil would all have boiled off by the time you have to replace the bulb.
There are specific "anti-seize" compounds, however, that're made for this task; get a compound that's made for use on fasteners in car engine bays and it should be able to handle light-bulb heat.
(A rule of Proper Mechanical Engineering is that no threaded fastener should be used "dry", because dry fasteners like to loosen off, or seize up, and it's hard to predict which. Anti-seize, or grease, or thread-lock "screw glue", should therefore be used on all fasteners, depending on their purpose.)
A dry anti-seize you might like to try is graphite. Just colour in the threads of the bulb with a lead pencil, and it should be much less likely to seize. Plus, graphite is conductive, while anti-seize compound probably isn't.
You can also get graphite powder in little puffers for blowing into mechanisms (and adding non-toxic blackness to the practical joke of your choice - toothpaste, talcum powder...). But honestly, a pencil ought to do it.
I was recently told by a ticket inspector who examined my newly not-working ticket (despite it working 20 minutes previous) that my mobile phone "must have zapped it". Being an engineer of the electrical variety, I was flabbergasted.
I was prepared to put this down to the one inspector until I looked at the ticket care website of my local bus, train and tram operator. Not only do they advise people to keep their tickets away from mobile phones, but also bud-style headphones. Am I going crazy here, or is this total baloney?
OK, so the ticketing system is based on 1987 technology, but surely both the magnetic field created by a mobile phone and the magnet inside of a bud headphone are nothing when compared with other sources?
No, your phone didn't zap your ticket.
The magnetic field around a cellphone, even when it's running at maximum transmit power, is way down in the milligauss. It will have no measurable effect on any modern magnetic data storage device.
(It might, just possibly, be able to damage data on 1950s-1960s magnetic core memory. But even then I think it could easily be an order of magnitude too weak, even with the antenna shoved right into the memory frame.)
It's barely possible that you could demagnetise a magstripe with iPod-type in-ear headphones, though. Even quite cheap headphones can be had with rare-earth magnets in them; those magnets are tiny and so the field strength falls off extremely rapidly with distance, but I can believe a field strength of a few hundred Gauss existing on the surface of the casing of some models.
It's usually rather lower than that, though, as evidenced by the rather weedy way in which the earphones stick to each other when brought together.
I love, by the way, the part on the ticket care page where they advise you to test phones for magnetism with a compass. A compass needle will have some difficulty responding to a magnetic field oscillating at the 900 or 1800MHz that modern mobile phones output.
(It'll still respond just fine, of course, if the phone has some sort of magnetic clasp, or for some reason includes a damn great DC solenoid coil. But a compass needle will respond strongly to a field much smaller than the few hundred Gauss you need to wipe a low-coercivity magstripe - the earth's magnetic field tops out at about 0.6 Gauss. Note that almost none of the "other sources" in the infoventures.com page you mentioned exceed 1 Gauss.)
After this page went up, a couple of readers reminded me that the speakers in most mobile phones have magnets on the back, which certainly will influence a compass.
Modern mobile phones are often so skinny that the outside of the casing is only a couple of millimetres from the back of the speaker, and magnetically shielding the speaker can make it too fat to fit in a slimline phone. So quite a lot of skinny phones these days can, at least, pick up a paper clip. (I just tried it with my current skinny Motorola and with my previous ancient Nokia; the speaker end of the slim phone can just barely hold two paper clips end to end, while the old Nokia could barely even swing a compass needle.)
It takes considerably more than a paper-clip-picking field to wipe even a low-coercivity magstripe, though. I suppose it's possible that some models of phone have rare-earth magnets on their speakers and can do it, but frankly, I'd be surprised.
Your articles about clicky keyboards sent me in search of a model M, and then I popped for a Unicomp Customizer, and then I finally came to the realization that what I really wanted was an old model F. Nothing feels as good under the fingers as a model F.
ClickyKeyboards says the converter box for model F's is no longer available. Jameco sold me an AT-to-PS/2 converter which doesn't work for the F (but DOES work just fine with the original model M I have).
Any suggestions where to find a converter? I've Googled until my eyes bleed and the best I've come up with is what appears to be an XT->rs232 electronics project. Now that you've rebooted my keyboard gene I've GOTTA have my model F!!!
(When I sent this answer to Dean, clickykeyboards.com indeed did not have an XT-to-AT keyboard adapter. As I write this, though, they have exactly one in stock, for a mere $US100.)
The sensible solution to this problem is to find an AT-interface Model F. I don't think they're common items, but these days they are, I think, more common than XT-to-AT adapters.
The adapter is not actually a very technically complex item; any microcontroller hobbyist should be able to knock one up for you quite easily, since the heavy lifting has already been done.
As you say, though, the adapters are also not off-the-shelf items any more.
If I were you, I'd set up a couple of targeted eBay searches, and tolerate the horrible agony of using a Model M until such time as a workable F solution turns up.
In letters #199 you write: "Note that the single highest-powered "extra" system in most modern cars is the air conditioner, but that's not a candidate for improvement here, because automotive air conditioning is almost always mechanically powered, using a belt driven by the engine."
I've been doing some reading about BMW's "EfficientDynamics", which trumpets the fact that it recaptures kinetic energy. I know that BMW had been looking at super-capacitors to store braking-energy for use when taking off again, but that's not what this is about. What it apparently is about is using engine-braking to charge the battery.
The more I think about this, the more clever it sounds. I don't know any details yet (if the alternator is only working some of the time, and if it provides a noticeable braking effect, is it then also bigger and heavier?), but it could well be enough to make electrical energy in the car virtually free in most driving situations. So headlights, electronics, heated seats, etc. no longer cost you petrol to run.
But taking the idea further, what about converting other devices that are powered by the motor to run off electrical power?
BMW have already replaced the power-steering pump by an electrical device - they claim that the electrical power steering is already lighter and more efficient than the hydraulic systems, but it's even better if electrical energy is free. I figure that air-con is probably the other logical target for this, although at some point I guess you might also need a larger battery, which will increase weight again...
BMW's Web site is less than totally informative, but if the Wikipedia article about EfficientDynamics is correct, all this part of the system is actually doing is disengaging the alternator when you're not braking and the battery doesn't need charging, but always engaging the alternator when you're slowing down. This will have some small effect, especially along with all of the other EfficientDynamics features. Auto Start-Stop, for instance, can have a large effect on fuel economy in city driving; it's the major reason for the great city fuel economy of hybrid cars. But by itself, the alternator trick is close to meaningless.
If you stop a one-ton car from 60 miles an hour, the brakes have to deal with the same 365,000 joules that it took to get the vehicle up to that speed (ignoring losses; we're driving in Physics Experiment Land, here).
A joule is a watt-second, so even if you stop the car over a somewhat leisurely ten seconds, the brakes are still turning an average of 36.5 kilowatts into heat over that period.
No battery that's smaller than a bus can accept that sort of charge rate, but you can dump energy at that kind of rate into things like flywheels, pressurised air or hydraulic fluid, torsional springs and other siege-engine-y sorts of contraptions. It's possible that capacitors will be usable in this role too; at the moment they can easily accept that kind of charge rate, but haven't the capacity to do so for more than a very brief time.
Then again, maybe I’m just too pessimistic.
It's actually uncontroversial that you can make hydrocarbon fuels from carbon dioxide and water. You just need lots of energy, to crack hydrogen out of the water and cook the CO2 into carbon monoxide. H plus CO is "syngas", to which you can apply the Fischer-Tropsch process to make all sorts of hydrocarbons.
The whole process is rather painfully inefficient, but it has the advantage that you can start running the existing fleet of petrol and diesel vehicles on what can be, ultimately, solar power. Pilot programs are in progress now; I saw one of them the other day on the third episode of James May's Big Ideas.
The Carbon Sciences people have the same goal, but aim to get rid of all the super-hot gas crackers and pressure-vessel reactors by using a "biocatalyst".
There's nothing ridiculous about this idea in theory - essentially, all you have to do is find some way to vastly speed up the process by which crude oil was formed. Whether their system is actually practical, though, I don't know.