Excuse a brief episode of second-hand blogging: something came across the wires that resonated with other things that have been posted here. Echoing my oft-stated concerns about the short-term nature of most digital forms of storage, the news is that a basic form of digital storage using very low-temperature semi-conductors has now been pioneered that, they say, should last 1000 years. (I got this from somewhere else, originally, but I’m gosh-darned if I can remember where, sorry to that brave resource.) The trouble with that, though it’s marvellous given that currently not much of our digital hardware can be guaranteed to remain usable for a twentieth of that, is firstly maintaining the low temperature—I mean, most things would last longer in perfect environmental conditions—but that knowledge often needs knowledge. One group of commentators at The Long Now Foundation have picked up on this, suggested that instructions for how to operate this device should be micro-engraved onto it, because that won’t be obvious in a few hundred years most likely. (Parchment and ink is still winning there, then…)
But to my mind that isn’t the big problem. The big problem is, about the only form of digital information one can store without encoding is plain text. In particular, you can’t hope to save pictures onto such a device: what format would you use that could be sure of being legible in a thousand years? Or is your chip going to have another chip with it with the Graphics Interchange Format standard on it? Although the lab in question is aware that really they need terabyte storage, the test chips so far only store a few megabits, though they will (and this is really cool) power up if subjected to a wireless interrogation signal without having to actually be plugged into anything, so the only interface you need is a wireless computer transmitter that’s been told how to talk to the chip. Though this is a tech. leap forward if it really lasts (and selling something like this as `fully tested’ is obviously not going to be possible), therefore, the basic problems of format and encoding are still there. Otherwise, the future may wind up thinking the art of the past was all ASCII unless they happen to have some really good quality paper prints or, well, manuscripts.
(That there, you see, is the monastery of Santa Maria de Serrateix converted to ASCII. Not really art-historical grade…)
The JPEG of Serrateix I fed to the ASCII art generator linked under both images to make the above
A Corner of Tenth-Century Europe 
It’s true, it’s a serious problem. The best solution so far is to only use non-proprietary publicly described formats for archival purposes so that given the information a skilled programmer can write a program to parse the file.
This of course assumes that we will continue to have skilled programmers and suitable hardware on which to execute the program, and of course assumes we all don’t go back to herding goats in some faux-medieval world as a result of some unspecified catastrophe.
There’s a school of thought that says we’ll have bigger things to worry about then, and there’s another that’s already inventing the clockwork PC :-) I have to admit that I was thinking, perhaps because of the mention of micro-engraving, of things like the LP on Pioneer 10 or similar messages to recipients quite literally out of our knowledge system. And that makes me think things like: what language do we engrave the instructions in… ?
what language do we engrave the instructions in… ?
Latin, obviously.
Touché! Though I suspect one of the Chinese languages might have a better claim…