Hacker News Reader

Comment by sdfjkl | Hacker News Reader

Amazing this stuff still yields secrets after thousands of years. Meanwhile the 3.5" floppy disks in my parents attic on which I've kept all my teenage Turbo Pascal source code have long lost all their data.

View on HN

Real programmers store their source code on baked clay tablets.

Ceramic is better. An interesting project here:

https://www.theatlantic.com/technology/archive/2017/01/human...

Nam shubs.

This is something I think about a lot. How do we build a "data pyramid" that can be interpreted by future generations. I feel like much of our culture will be lost to bit rot.

Wikipedia could be encoded into DNA and them spliced into species that are very long lived, geologically speaking, such as horseshoe crabs. Absent any negative selection against "wiki DNA", they could remain readable in millions of years time.

You will need a very strong error correction scheme. It is not necessary to have a negative selection against "wiki DNA", if there is not positive selection the error will accumulate quickly.

With a quick googling, https://en.wikipedia.org/wiki/Mutation_rate#Variation_in_mut... I think that a mutation rate of 10^-5 per base per generation is a good estimation for humans. Let's assume 10^-6 per base per year. So after a million year, with no selection pressure, probably more than a half of the bases will be overwritten with garbage.

Losing only 50% of the data after a million years doesn't sound that bad. The bigger problem is keeping knowledge that the data is there and how to get it to survive even a tenth of that length.

Probably there would be "signal decay", but one can still make sense of a text when many of the letters are missing. So the decay would have to be more than 50% for the message to be completely useless.

Also, a million years is quite a long time, I doubt there are many media one could use that one could expect to still be at least half readable in one million years time.

It'd be smarter to simply integrate it into a genome, make copies, extract it, then lyophilize and store it. Don't house it in an organism.

I guess my issue with this line of thought is - how do you include the "playback device" along with the media?

So many assumptions must be made about the theoretical future inhabitants that would "discover" our culture would be capable of technologically.

How do we send the VHS tape, along with the player, instructions for how to setup the player, play the tape, output to a sensing device of some sort, and only _then_ you can possibly read only part of the data if you get that far?

I don't think I will even have photos of my kids or anything from about the last 10 years or so by the time I am in my mid fifties (not as far away as I would wish).

Edit: mobile make bad spells

There are 2 of these projects I know of in existence. E.g. https://www.theatlantic.com/technology/archive/2017/01/human...

Also with a lesser extent of knowledge storage this one is interesting: https://www.wired.com/2009/04/ff-guidestones/

One line of thought I was going down recently was information encoded by printing very specific and precise three-dimensional "monoliths" that display information based on postion of the nearest star (which I assume would still be Sol, and would still be around in this distant future).

To give an illustration of what I am trying to describe, I am imagine something like the cover of GEB (https://images.gr-assets.com/books/1331382003l/24115.jpg), but on a multi-dimsensional scale.

Full disclosure: [7] ;)

Nice idea for short term (galacticly) but as far as I know the orbits of the solar system (even the 'stable' ones[1]) are unpredictable (chaotic) to a certain degree over long timescales. Assuming sol doesn't screw us over and the output behaves as we expect.

[1] https://en.wikipedia.org/wiki/Stability_of_the_Solar_System

True. I suppose my point is that I belive that a _physical_ artifact is the only chance at this working. What if the culture that discovers it is in a "dark age" and doesn't even have the electricity to power these supposed future proof digital archives.

Digital preservation is the name of this field. It's pretty interesting.

Way off topic, but have you actually tried to read them? I was able to get data off some ~25 year old 5.25" C64 diskettes a few months ago, encouraged by success with data recovery from some 3.5" diskettes a few months before that.

Yes, most of them didn't last more than a few years before the bits started falling off. This may partially have been because most of them were SD (720kb) that I'd drilled the HD (1440kb) marker hole into to save money. Worked fine at the time, but probably didn't help boost their longevity.

Reply on news.ycombinator.com