There are CDs, DVDs, Flash Drives, Hard Drives, Floppy Disks, etc. that we can backup our information on, but none of these are guaranteed to last more than a few years as far as I can tell. If I have extremely valuable digital information, what’s the best medium for backing it up so I won’t lose it in the event of a hard disk failure? I’m talking 20-30 or more years.
And as a corollary question, why isn’t there a consumer level archival medium? Is this really that hard a problem to address? And why aren’t flash drives reliable over the long term?
Yeah, most places say the best thing to do is have copies of it on multiple sources (HDD, DVD, tape, say), and regularly review new tech, and copy over to the new format when it seems solid.
All sounds kinda lame and a lot of work, but truth is, there’s little progress in this areas, especially at a consumer level.
Properly stored, CDs and DVDs should last 50+ years. If you expose them to a ton of sunshine or humidity, they’ll degrade, but kept in cases, in a shady place, they should last a long, long time.
In truth, there is and can be no future-proof, maintenance-free storage medium for electronic data because the common formats for that data are subject to constant change, as is the hardware upon which the applications to manipulate it are implemented, but storing in machine-readable format means that porting to a new format will be easier when standards start to shift. Storing as any kind of hard copy just introduces two more stages of transcription to the end-to-end process.
This might not matter if the data is naturally human-readable - i.e. the complete works of Shakespeare, where there’s no pressing need to have it presented exactly the same way every time, as long as it’s recognisable text - but for something like a high-resolution, high-colour-depth photograph, hard copies are quite likely to be lossy, certainly so over time - defeating the object of archiving in the first place.
If your data is truly valuable, about the best you can do is put it on whatever media is available, then annually, read the data off and write it onto a new whatever. This will let you stay current with media formats and avoid long-term deterioration. CD-R (not -RW) is your best bet for something that will be in use five years from now. There are quite a few DVD formats duking it out and it would be hard to predict which will still be in use in 2010 and beyond.
Also, archive at the lowest-common-denominator (or native) format, as well as whatever form it’s currently in.
ie: archive Photoshop files, or even JPEG images as TIFF images in addition to the application’s (possibly proprietary) format. If your digital camera uses RAW, that’s also a good choice for archiving. Likewise, any words should be saved as .txt plaintext.
Databases and spreadsheets can be saved as tab-delimited text files. An older format such as Excel 5.0 also stands a better chance of being readable by whatever application exists in the future.
A claimed 300-year media life is all well and good, but how many applications and data formats have come and gone in just the past 20 years?
As one example, I had to toss out a couple dozen graphic elements that I’d created in version 2 of an application way back in 1993 as nothing newer than version 4 can read them. Nothing newer than version 7 can read version 4. My only other option would have been to scrounge up version 4, open the files, save them in version 4 format, then find version 7, read and save, then hope current versions will continue to be able to read version 7 for a while. It was far quicker to simply re-create the graphics.
This app has had two obsoleting format changes in 12 years. Care to place any bets as to what the next 12 years will bring?
Punched mylar tape has a very long shelf-life and excellent reliability. It is also easy to build an optical tape reader from scratch. The only problem is its low data density.
My solution is to keep a workable driver available. Watch for hardware dock shifts, at the outset electronic stores such as Radio Shack® usually carry adapters.
This has worked fairly well. Application problems are another matter.
Well, you couldBinHex the digital file (regardless of what kind of file it is) and then open the .hqx file in a plain text editor and print it out at a moderately high point size on acid-free paper.
That ought to be very durable. Two generations from now your descendants should be able to scan those pages in and run an optical character reader on them and decode the resulting text file back from BinHex to the original binary.
The author strongly makes the oft-missed point that for most non-trivial data, the ability to read the bits off the storage device is the least of it. Without the software to interpret the bits, and the OS to run the software, and the hardware to run the OS, the bits are effectively undecipherable, even if they’re not encrypted in any way.
