Very Long-Term Backup
kk.org
kk.org
Patek Philippe may prattle on about how their watches are meant to be passed on to descendents, but this is something whose use is explicitly dependent upon outliving you. Heck, losing it adds to its value!
I'm not one for knick-knacks but I'd love to own one of these. Even as a fad, profitable sales could considerably increase their number and geographic distribution while reducing their cost (hand made at $25k? There must be a way to automate that!).
Or you can have one on earth, if you want, acting as an additional node in the distributed archive. There are still two disks available from this prototype run. Currently, for all its high techness, each disk is hand crafted, and so they have a corresponding high hand-crafted cost: $25,000. Contact the office if you are interested in caretaking an archive of all languages.
The currently accepted solution is to copy everything up to "the cloud" where things are backed up and redundant. While this is a good solution for a lot of people, it's not for everyone. I get the feeling that there's a group of (mostly older) people out there that would want (and be be willing to pay for) a longer lasting, physical storage medium like this. A long-lasting write-once medium like this might be a good compromise between digital (easy to store a lot, fragile) and analog (harder to store a lot, longer lasting).
The main issues with this type of medium, though, would be price and compatibility.
From what I can tell, this is also in the creator's best interest as that is the natural direction for this project -- One in every household.
Most of the translations are translations of translations, not direct (and some are even 3rd generation), which will hurt analysis if you don't know which the original is.
Do guaranteed-safe data storage right and you will have a killer app. How do you really plan to last for 100 or 1,000 years? We all have lots of value tied up in our data. Think of long-term storage as time-shifting that value, just as files are a way of time-shifting data processing. I believe cperciva is working on this.
The real problem here isn't backups. It's archaeology. Keeping the data you want is easy - just about everyone in computing is working on some version of that.
How are we ever going to get that one back when we need it?
A bet that google will still be around in 100 years, or will buy insurance to guarantee long term storage via some foundation if they do go out of businesss, seems reasonable to me.
GFS does CRC and multiple copies to guard against data corruption right?