I copied backup data to an SD card, then I placed it at the little plastic container, then the button on my pants turned it into sd card pieces when I sat on a metal surface.
It occured to me that if the SD card had multiple data paths (gold lines) on the outside of the plastic as well as a system that automatically copied the files to different corners then even seeing the bare silicon the data could be refound
also this provides a way that 64GB sd cards can be offered as 16GB rugged super durable versions providing some benefit to sd card makers
also it occurred to me that if the came up with a new USB standard that was like 1V they could make USB thing n data cards with conductive polymer leads. It looks like a 3$ multiGB sd card could actually be ten 30c or cheaper data cards for things like music or media DMA (paper correspondence) so Greenpeace could send people movies along with the paper letters.
The actual memory portion of an SD card is cheap enough, its just changing the metal USB part compared with a new specification that makes an most all polymer USB cheap as a CD to create the technology
note: you can draw 1v with the new all polymer USB specification then use an all semiconductor voltage multiplier to power the flash memory-- beanangel, Nov 29 2011 The USB CD The_20USB_20CD [beanangel, Nov 29 2011] just no ... Even if you have the same data all over the sd card, you'll have to take it to a specialist for data recovery to get your data back when you shred it. You could do the same thing with a normal sdcard and it will cost a heap of money in both cases.
Also the price to space won't scale linear cause you have some control logic that will have to be duplicated in every sd card. Actually i think that the manufacturing costs of a 16gb and a 128gb sdcard are almost the same.-- fho, Nov 29 2011 I agree with [fho] that the price scale won't be linear, but I disagree that you could get the same benefit with a single SD card. First is that a single SD card probably has a single silicon chip. If you put a large force on one side of the chip, you are likely to cause damage all over the chip. If you have multiple chunks of silicon, each in a plastic package connected with contacts, one device can be completely destroyed without stressing the others. Second, if the system is designed properly, there would be no need for special equipment to recover data, whereas recovering data from part of a chip that is damaged is a very touchy operation.
One fairly simple implementation: embed 2 Micro-SD cards in a full sized SD card. Add a controller chip to make it write the same data to both Mico-SD cards and give a read error if the data from the two cards doesn't match. When the SD card gets crunched, the user carefully opens the SD card shell and removes the two Micro-SD cards. If it isn't obvious which one is broken, they can try to copy the data off both using a standard micro-SD adapter and see which files look right.
If you want to have 4 copies, you could make a proprietary package type that is smaller than Mico-SD, and include four of them in the SD package. A "recovery kit" that was just an adapter to this proprietary package could be sold separately or included with the original device.-- scad mientist, Nov 30 2011 So, you're proposing not having one, large, fragile chip but having 4 small (and hence harder-to-break) chips - why not just have one small chip?-- hippo, Dec 01 2011 While it is true that a smaller chip is generally more robust than a large one, there is even more protection offered by making it redundant. That protects against some additional damage scenarios as well as providing protection against failure of the one of the sub-devices due to age, etc.-- scad mientist, Dec 01 2011 random, halfbakery