Dan's Data letters #67Publication date: 19 October 2003.
Last modified 04-Feb-2014.
The county I work for is pushing digital imaging very hard as the answer to document storage. We have a warehouse full of archived documents, and they don't want it to turn into warehouses (plural). My concern (which they don't seem concerned about at all) is the longevity of the digitized data. See, for example, this article (unofficial English translation here).
I was surprised to see that they had CD-Rs with high failure rates after just 20 months. Actually, I wasn't that surprised, because I've been hit with bad CD-Rs, too.
So just what is the most robust, and yet economical way, to store digital data long term?
This got chewed over on Slashdot a while ago.
The PC-Active results are strangely lousy; maybe they didn't store some or all of the discs as well as they should have. The take-home message that CD-Rs shouldn't be considered a long term archival storage solution is still sound, though.
There's no single, bulletproof, long term data storage method (well, barring Rosetta Project disks, or something). You face not only media deterioration problems, but simple compatibility issues. 20 years from now, it's quite possible that you'll have to go to a special bureau to find a CD/DVD drive.
This isn't actually a problem, though. The steady march into larger and larger capacity storage means that every few years, you should just take a weekend to copy all of your old data onto whatever format is now the new hotness.
Today, that means reading all your old CD-Rs and copying them to DVD-Rs (which should drop the disc count by at least a factor of six) and to commodity ATA hard drives ("120Gb" drives, with around 114Gb formatted capacity, are the value leaders at the moment; more than 165 full 700Mb CDs per drive...). Make (at least) two DVD-R copies and two hard drive copies of everything, store them separately somewhere cool and dry where they're unlikely to get burned, stolen or drowned, store the old CDs somewhere as well (because, hey, why not, if you've got the space), and you're done for the next five or so years.
Then, in 2008 or whenever, you take another weekend to copy all of the old stuff, and everything you've accumulated since, to Blu-Ray discs and 1.2 terabyte drives, or whatever we're using by then.
Probably, each copy operation will take about the same amount of time, cost about the same (which is to say, not much from the point of view of most businesses), and require about the same extra storage space. Since you should be reading your old backups every few years anyway to make sure nothing's dying, the incremental cost is actually quite small.
A co-worker has been religiously cutting pins off all his old CPU's thinking he can make his girl friend a nice trinket from the gold. Is he on crack?
Well, if he is, that could explain why he can't even afford to buy a necklace on eBay.
Gold contacts on various pieces of computer hardware aren't solid gold - they're plated. Gold is used because it doesn't corrode. Yes, you can salvage gold from old computer components, but it's not really worth doing unless you've got a factory in some benighted nation where it doesn't matter too much if a worker falls into the nitric acid bath.
Here's a Usenet post on the subject.
My digital camera is really slow saving TIFF to its memory card. Is there any advantage to using JPG to the card and converting to TIFF later on the hard drive?
The best format to save in, if your camera supports it, is RAW; that's the unprocessed sensor data. TIFF is a fairly distant second; it's been "de-mosaiced" into a normal full colour image, rather than the matrix of single-colour pixels that the camera actually sees. Next comes JPEG, which even at high quality settings throws away significant data.
The only reason to convert JPEG to TIFF is if you're planning to do post-processing. Every JPEG load-and-save cycle will damage the image a little more; TIFF is lossless, and you can load and save it as many times as you like and not change it at all.
For non-critical shooting, moderately high quality JPEG is fine. I almost always shoot JPEGs myself, because I'm usually just doing product shots, not landscapes or portraits or anything else really important.
I saw the "CPU on a wire" letter [on this page]. Then I thought about the increased heat from the CPU and video card. Why not turn the CPU socket around and have the cooler sticking out of the (tower) case? That would separate the hot air from all other components. Would that work?
This'd be physically possible, and would certainly allow better CPU cooling, but it'd hardly be a marketable product. You'd need a hole in the side of the case, of course, and now you'd have a computer that'd be pretty much utterly destroyed if you bashed the piece sticking unexpectedly out of its side.
It's more sensible to fashion a separate CPU cooling duct, on the other side of the case. Many people have done this, some elegantly and some with cardboard and an axe; it works pretty well either way.
I've noticed in the last day or so that the clock in my computer is running fast. I mean, really fast. It's gained about 3 hours today. What gives? What could possibly have caused this to happen, and is there anything I can do about it?
This problem could be as simple as a dying motherboard backup battery; these days those batteries are almost always simple lithium coin cells that're easy to replace.
It could also be an OS time server sync issue. If the computer keeps good time when it's turned off, but gains time when it's on, then the mobo clock is fine and the OS is screwing up. Windows XP automatically does periodic time syncs with Internet time servers (defaulting to time.windows.com; look in the "Internet Time" tab of Date and Time Properties) if it's got a Net connection; other Windows flavours can easily do the same thing with the "net time" command or extra software (the shareware YATS, for instance, or the free TrueTime WinSync).
When you sync to an NTP server like this, your computer can get not only a time setting, but also a time skew setting - a correction factor that matches the computer's idea of the length of a second with the server's presumably more accurate idea.
If your computer's come up with a wiggy skew setting, that can cause it to gain or lose time drastically.
Just re-syncing the time, possibly to a different server (try the second option, time.nist.gov, if you're running XP), may cure it.
Here's a Usenet post on the subject.
This page pretty much covers it. RSS gives you an easy way to keep up with updates on numerous sites, without visiting them over and over, or just using some dumb "site has changed" notifier that doesn't tell you what has changed.
If you're looking for a simple RSS aggregator, by the way, I think Bloglines is pretty cool.
When you RMA a set of Z-560 speakers (in Canada, at least), Logitech only require you to send in the control module, thereby leaving you a set of ostensibly useless satellites and a sub.
I was considering using the satellites for the rear and surround channels of an inexpensive home theatre system. I believe if I disassemble them and solder in a 100Hz high-pass passive crossover, they should work reasonably well.
I also think I could disassemble the sub, solder in a coil, fashion a new back side out of MDF to replace the electronics, connect spring clips and end up with a decent car audio sub.
Any flaws you can spot?
The satellites would probably be fine as surround speakers. You wouldn't need much of a crossover, either; just putting a 220uF cap in series with them would probably be fine.
The subwoofer idea would work too, except the Logitech driver is an eight-ohm unit, so you might have trouble getting enough volume out of it from a car amp that expects four-ohm speakers.
For the last year or so I've kept an elderly, Socket 7-based desktop PC (one of those phonebook sized things Compaq used to make) tucked behind the seat of my pickup truck. I swapped the standard 200MHz MMX Pentium for a K6-2 400 and it now happily serves up MP3s, does GPS navigation, and occasionally serves as a platform for wardriving (um, I mean for conducting wireless Ethernet security surveys). The display is an older, but still serviceable 10.4 inch LCD Keycorp monitor with matching PCI video card I picked up for cheap on eBay. A small, IR wireless keyboard with built in pointing device provides the user interface. The monitor is powered from the video card, while the GPS receiver and wireless NIC are both powered from their respective USB ports. A wireless FM modulator pushes the sound card output out to the car stereo system (and to those of my friends or anyone else who happens to be within about a 500 foot radius) I figure the total system, under full load, probably draws somewhere in the neighborhood of 50 to 60 watts.
All in all the system performs very well and cost me next to nothing. My only real complaint is that if I turn off the ignition, I have to shut the PC down before I can restart the truck. You see, I'm running the system off of a small DC/AC inverter and when I crank the engine, it senses the large voltage drop and immediately shuts itself off. This makes short stops for food and fuel very annoying. I tried adding a small UPS into the loop, but it absolutely refused to accept a charge from the inverter. I figure the little electrons protested being converted from AC to DC and back one more time and decided to just pick up their lunch boxes and go out on strike. I believe that in my area, electrons are represented by the United Brotherhood of Subatomic Particles, Local 9.1093897x10^-31.
Is there any (cheap) way you can think of to power the system during the 2 to 3 seconds it takes to start the truck?
Hmm. Monstrous capacitor? Car stereo nuts use one-farad-plus ultracaps to hold their power rail up and stop everything dimming every time the subs go WOMP.
Given the length of the starting process, though, I'm inclined to think you'd need a second battery and an isolator, as used in recreational vehicles, boats and such to prevent the fridge/TV/Playstation from flattening the cranking battery.
Here in Australia, Jaycar will sell you an isolator kit for $AU49.95 (stock code KA1782). Places that sell RV equipment should have pre-built ones; a basic fairly low-current model would probably be more than adequate.
Any old car battery would do as the second one - actually, a little "gel cell" would probably cut the mustard, as long as the isolator didn't dislike it.
Hello. One quick question about your Those Darn Wires column and microwave energy transmission. Couldn't narrow high energy beams be used in a tunnel? I have no idea how wide such a tunnel would need to be, but it would be safe enough I imagine.
You certainly couldn't just stick a giant magnetron at one end of a disused iron gas pipe and expect to get something worthwhile out of the other end, but yes, it is physically possible to make a calibrated waveguide conduit that can pass large amounts of microwave energy. It's less easy to efficiently convert your electricity into microwaves at one end and back at the other, but let's handwave that for a moment.
The power transmission capability of a waveguide is limited by the dielectric breakdown voltage - the insulation capacity - of whatever's filling it. Air's a pretty good dielectric for everyday purposes, but it breaks down just fine at very high voltages, as you may have noticed during thunderstorms. The result of dielectric breakdown in a multi-megawatt transmission conduit would be spectacular, to say the least; internal arcing will result in all of the energy being reflected back to the source, which, depending on the number of feet of armour put between the monstrous magnetron and the power facility staff, might then result in the power company sending a large number of condolence messages to their employees' families.
If you evacuate the conduit then you avoid such problems, but keeping all of the air out of a conduit many miles in length would not be cheap.
Is there such a thing as a 1.5 volt NiCd or NiMH cell, in particular a AA, and if not, is there any reason to do with the chemistry of the batteries that would stop someone from making one?
I realise that they don't have the same voltage drop as a chuck-away battery, and therefore 1.5 volt versions could damage some equipment, but surely this isn't the only reason you don't see them around.
Perhaps they could make a compromise, a 1.35 volt NiCd/NiMH AA for instance, surely that wouldn't harm anything?
NiCd and NiMH cells are all 1.2V nominal - a bit more right after charging.
The NiCd/NiMH chemistry inherently has lower volts per cell than carbon-zinc and alkaline manganese non-rechargeables. This isn't something that battery manufacturers can change; a given electrochemical cell always develops exactly the same potential.
When I recently moved house and found the place full of vertical blinds (the ones you find in offices with two cords, one for opening and closing, and the other to rotate the "blades") I immediately found it irritating trying to remember which cord does what. I can see this annoying me for years to come. The solution? An electric motor gizmo that is remote controlled. Simple. But not yet invented it would seem.
Does such a thing exist, and if so where can I buy one?
What you're looking for is referred to by home automation geeks as a "drapery control system". Most of these things seem to deal with only one action - drawing curtains or pulling one blind cord - but there are plenty of dual-action ones as well.
I am tired of recieving pointless emails from readers on your site [specifically, the end of this column]. I don't quite understand why my email address and my home address have been put onto your website. This is out of order, besides it says you do not collect information about people and do not publish email addresses on the privacy section on your site. Your friend may have told you the reason I behind breaking a computer to get a new one(somemoney off a new one in fact). When I bought the computer I said I didn't want the the warranty, the sales woman however told me if I get it I will be able to break it when it becomes old and they will give me a new one. She told me that the store manager had done the same.
Im pretty sure you would have done the same.
You're tired of being e-mailed by people who think you're a dork, I'm tired of being e-mailed by people like you who want me to help them commit a crime; I guess we're even, eh?
Your home address is not on my site. Your address is in your domain registration information, which is public information; I just linked to a whois gateway.
If you can't figure out why I'm angry about being asked to help you commit fraud, I suggest you think a bit harder about it.
It is, of course, discourteous of me to invite people to tell you their opinions about your behaviour. It is my opinion that in asking me to help you commit a crime you have forfeited the right to be treated with civility. If you disagree, I invite you to sue me.
If, as you say, the saleswoman at the computer store told you in so many words that you should break your own computer to fraudulently claim on the warranty, then you already have a co-conspirator in your attempt to commit fraud. I suggest you contact that saleswoman again, and ask her how she thinks you should do it.
And no, I wouldn't have done the same thing. I wouldn't have bought anything from a company like this. Obviously, it's at least staffed by crooks, and may be run by them too.