By all means preserve what you can, but I'd like to relate a personal anecdote.

I'm an old guy, born a little after WW2. My father was a child of the depression. During WW2 he was trained to be a photographer's mate in the navy. Afterwards he worked in civil service, but moonlighted as a photographer, usually a wedding photographer, during my childhood years. He was very good at it. (Maybe that excellence of training is how we won the war.) Later he taught me both photography and darkroom techniques.

Just once I accompanied him on a wedding shoot, when he had injured his hand and needed someone to help carry equipment, and I learned something about the paradox of wedding photography. In a formal wedding, the photographer is the second most important individual perhaps after the clergy, but has a lot more effect of the proceedings. The wedding photographer must create and preserve a permanent stylized record of the _perfect_ wedding, while not getting in the way in the process of the _perfect_ wedding. That means staging and composing the _perfect_ record of the wedding party, and the various stylized posed events of the ceremony, while not leaving anyone with the perception that the entire event is being staged entirely for the purpose of its photographic record. This is difficult to achieve.

I still don't know how he accomplished this with a young, tense wedding party, but accomplish it he did. My father was not well educated (I, on the other hand, have been a college professor) but he sure knew a lot. I can only explain to myself in that I did not live through the depression, so I cannot understand.

But anyway, I can offer this advice. Since your children are old enough to remember what will happen during your wifes remaining lifetime, let them remember that as naturally as they can. By all means digitally record what your can, but don't let the process of that recording distort the experience they can have with their mother during her remaining time. Achieving this is a skill you don't have opportunity to learn, so when you are unsure, let life happen instead of recording it. They will remember!

Best wishes for all of you.
 

I agree with those for whom the always online requirement seems excessive. But that is what the OP wants. Here's an idea how to do it with minimal cost.

Buy 16 or so 1T USB drives which can be had for less than $100 these days. (The drives need to be self powered.) Buy 5 quad-port USB hubs along with any necessary USB cables. Plug the 16 drives into the four hubs, fanout those four hubs from the fifth, and plug that last hub into the computer. Variations are possible with different fanout, or spreading two top-level hubs plugged into the two separate USB channels on most machines.

Now, I have no idea whether this would really work, but is something I've always wanted to try. (Although I have no actual need for it.) The USB specification is supposed to support multi-level fanout, but I've never needed to try it. Anyone know why it wouldn't work?

One key to this idea is that, unlike datacenter servers, the OP probably needs only serve a single large file at a time, so bandwidth requirements are modest.

Since you're interested in radio, whatever other things you find I suggest strongly that you google up a local ham radio club. Not all hams do circuitry construction and tweaking, but you will surely find some who do and who can be valuable resources for advice, tools, component sources, etc.

There is a whole subdomain of ham radio that does digital packet switching, if that piques your interest.

It's been about 40 years since my General Class license last expired, but 100W rf emission does not compare to the many times that power which would be necessary to power the equipment necessary to send that effective power into the ether. I assume nowadays that semiconductors would reduce the overhead, but (for instance) back then it was necessary to heat the tube cathodes bright red before they would emit significant quantities of electrons...

"Damages" in a civil complaint, despite legal theorems, has two components. The obvious one is the the costs suffered by the plaintiff. The other (less often explicitly acknowledged) is the punishment to discourage future repetitions.

Suppose some extremely-clever human-engineered phish or spam yields on average more than the fraction of a cent cost that span penalties might obtain. There would be no disincentive for spamming

Of course, spamming today has essentially no cost to the perpetrators. When there is an international corps that track down spammers and either puts a bullet in their brains or shuts off internet connectivity to their entire country, operating characteristics will change.

> My blackberry has a microSD card in it. I have passed through many different customs / airport
> security examinations and nobody has ever examined the contents of the card. I don't see the
> point of paying for an even smaller microSD card carrier, when I already have a small microSD
> reader that I carry with me everywhere that nobody ever raises an eye towards.

Certainly so. The ways of hiding information inside a piece of digital electronics are too numerous to enumerate, and it is far easier to transfer huge amounts of contraband data safely using various internet protocols.

But suppose one absolutely needed to transport something like a micro-sd card through border control. Your completely innocent cell phone or laptop or personal dvd player is full of electronics that look to a X-ray very similar to an sd card. Just find a cluttered place on the motherboard and tack the micro-sd card with a little rubber cement or tape. Be careful to preserve the alignment (most chips are aligned at right angles to with the motherboard) and the chip will be unnoticeable to X-ray. And since it is completely unconnected to the mother device, it will also be undetectable under any software exploration of the device.

Subterfuge and terrorism are easy. Policies that make ridiculous the motivation for subterfuge and terrorism are far more difficult.

Regardless the motivation behind this project, if I wanted to experiment with using a disk to store something other than regular data "blocks" I would try experimenting with an old 3.5" floppy rather than a modern hard disk. (Also, I'm incredibly cheap.) Old floppy drives and media are available for pennies. Control of the device, and especially formatting, is much closer to the processor. (There wasn't much smarts in floppy drives.) One could perhaps find formatting code in some early Linux distros. It might be that the proposed project needs to use a real HD made out of real metal, but this might be a way to experiment. I concur with earlier observations that there is so much smarts in a modern HD between the computer and the drive components that making the drive work without all that stuff would be daunting.

I'm no expert on criminal justice, but Congress abolished parole from the federal penal system in 1994. See http://en.wikipedia.org/wiki/Parole That means barring successful appeal or very unusual intervention (e.g. presidential pardon) when one is sentenced to a term in federal prison, one does that term.

Last week I overheard some Googleheads at the bar. They were talking about a new line of business for Google that sounded really neat. Customers could order groceries over the web and orders would be filled at highly-automated local warehouses and then delivered by GoogleVans. Wonder why Amazon didn't think of this first....

Chrysler as well. Some time in the late '60s a friend's father gave him a dead 1948 Chrysler -- can't remember to model. It has an inline 8 and would no longer start. We towed it about 50 miles from a D.C. suburb into the city, and the Chevy doing the towing in the summer heat never again ran very well. But when we arrived, the Chrysler magically started! That was a great car. It had those neat external visors on the windshield, and one could stretch out one's legs fully in the rear seat. The 8 cylinders under the hood were in a slanted straight line. That vehicle must have been about 37 feet stem to stern...

Back in the early '70s our favorite computer (PDP-1 serial number 3) had a large-format drum plotter that could draw and print pretty fast. The drum moved the paper vertically, there was a carriage holding a pen that moved horizontally, and the pen could be raised or lowered by a solenoid. It was the only practical way to make hardcopy of anything that wasn't text. After a while one of the locals rewrite the block-letter printing routines (mostly used for labeling, etc.) to write in cursive on the theory that cursive was already optimized for speed, reducing the need to raise, move, and lower the pen. He was right, although the difference was not large. But I suppose this experience does not much apply to current raster devices... Anyway, this was a neat hack, long before ascii video terminals or raster devices were commonly available. More seriously, my other line of work (in the humanities) suggests a continuing importance for retaining cursive. There is a long tail of historical documents in cursive. If you have a cardboard box with some of your great grandfather's letters, they are unlikely to be typewritten. If you need to research historical documents and letters, you will need cursive. Indeed, in a research library I once encountered a couple 19th C cards in a card catalog that were in cursive. (Of course, card catalogs themselves are now mostly gone.) German Handschrift for a couple centuries leading up to the early 20th century is mostly opaque to me. I wish I could read it, but never put in the effort to learn it. It would be a shame if most of our population could not read our original documents that might be less than 100 years old.

It is curious that no one has noticed, or at least commented, that such a battery could be dangerous if not used properly. The standard EE model for a battery is a voltage source in series with a resistor. This is called the internal resistance of the battery, and its effect is to limit the current that can be delivered by the voltage source. All batteries have nonzero internal resistance, and for most batteries, that resistance is significant. There are many kinds of batteries that you can safely short, and nothing hazardous will happen as the battery discharges as rapidly as it can. In order for the battery described in this article to be able to deliver such significant power in such a short time, if it obeys the simple first-order model, it must have an extremely low internal resistance. That means that if such a battery is shorted with a conductor sufficiently robust not to vaporize, the current flow through the shorting conductor could generate sufficient heat so as to be a serious fire hazard. There are videos on youtube showing what happens to a laptop when its battery is artificially induced to discharge catastrophically. It isn't pretty, especially if the laptop were in your car trunk, your backpack, or under the seat in a commercial airplane. I remember back a zillion years ago when a roommate's car was frontally crunched. He decided to loop a chain around the front grill and mounts and try to pull the bodywork into shape using another car. The chain cut into the insulation of a battery cable (probably the started cable, which mysteriously was routed along the radiator mounts) which shorted the lead-acid 12V battery to the frame. A wreath of smoke emerged as the insulation on this rather substantial cable melted, burned, and vaporized. Fortunately the now-completely-junk vehicle didn't catch fire. I believe the internal resistance of an automobile lead-acid battery is somewhere less than 0.1 Ohm. Consider what the internal resistance must be for the battery in the article...

I haven't read through all of the myriad replies, but the repeated themes of the responses illustrates that asking a design question on Slashdot will yield predominantly technohead solutions. The original query is rather unspecific about the problem it is wanting to solve. Why is it necessary to _bury_ the data, and why is it necessary that the data be in _digital_ form. Presumably the real requirement is that the data be recoverable after 25 years. (Aside: 25 years is a rather short time for data preservation, and extremely short for a traditional time capsule. I've been in computing nearly twice that long. But I've noticed from anecdotal new reports of opened time capsules over the years that the capsule has usually failed mechanically over the years, and that stored items are compromised by moisture leakage.) Anyway, thinking back 45 years, I can't think of any medium from back then that would be conveniently readable today (except scanning carefully-preserved paper). Even thinking back the specified 25 years, almost any medium would not be readable, although IDE drives existed around then and a hard floppy disk would just barely qualify. But why not think outside the hardware box? How much are you willing to spend? Although it is questionable whether media can be preserved 25 years, there is no problem with media that is only 5 years old. I can think of no current, main-line media format from 5 years ago that could not be read today. So why not create a trust that will rent several redundant geographically-distributed bank safe deposit boxes, and store multiple copies of your favorite medium, or several favorite media, in each. The trustee (which could be a bank or other financial institution -- it doesn't matter if they go out of business, because courts treat trust responsibilities very seriously) would be required every five years to engage a consultant to decide whether and how the media should be recoded and restored in light of current technology. This could be established with a reasonable endowment that could return an inflation-adjusted $10K per five years, to cover both the trust and consulting fees. If the banks, government, and Western civilization all fail, then perhaps you won't be concerned about your images. But otherwise preserving them for 25 years is easy, if you can pay the price. The price falls greatly if many clients take a share. Perhaps this will be a startup suggestion for someone.