Hacker News Reader

Comment by fsloth | Hacker News Reader

I think we are nearing the point where someone will come up with a solid format storing bytes, in known encoding, for a quite long time. The incentives are certainly in place - libraries specializing in data storage (e.g for hollywood film master copies, etc), are to my understanding spending a considerable buck in updating theis storage scheme every so few years. Despite how many megahours of youtube data gets uploaded daily, I'm fairly sure there is a market for relatively low capacity but longterm viable storage once someone invents it.

View on HN

There is such a thing as 5d memory storage, which I understand basically uses etched glass and is thus super mega stable, although I'm not too sure about the costs involved in heavy use for things like day to day archiving https://en.m.wikipedia.org/wiki/5D_optical_data_storage

A market, maybe, but do you think it will be a big one? Where is the incentive to spend a lot of money on this problem?

Existing libraries already have trouble with low-quality books and film that doesn’t survive very well. Only the most important artifacts tend to be copied to archival-quality material.

Why would digital artifacts be any different? And where is the incentive to preserve this stuff? Look at how much bitrot there is on most web pages after just 10 or 20 years. Even setting aside physical bitrot, there’s plenty of semantic data loss that nobody really cares enough about to fix.

Finally, despite all the above, let’s say we do come up with a cheap and practical way to store digital data for very long periods of time. Unless it’s very compact, it will use an appreciable amount of energy and material resources to store stuff. Which tiny fraction of the massive tidal wave of digital content should be archived, and who gets to decide?

"And where is the incentive to preserve this stuff?"

To my understanding there are institutions including the library of congress whose main purpose is preserving information. Not all information, of course.

Reply on news.ycombinator.com