[astine]

These kinds of technologies have been tried before (and largely abandoned.) The real problem in long term data storage is figuring out how to interpret decades old binary formats for which the software no longer exists. It doesn't matter if you have a thousand perfect PDFs if Adobe has been out of business for a hundred years.

The other issue with long term data storage is that the whole advantage of using digital technologies in the first place is that data is easy to copy. Using specialty technologies like this totally nullifies that advantage.

If your goal is simply long-term data storage/archival, you'd do better to just use durable and popular media of the day, to maintain an index of your collection, and to upgrade/migrate your media and file formats regularly to keep up with the times. If you're going for really long term storage, such that might go multiple human lifespans without being needed or needing maintenance, you might as well just skip the digital to begin with.

[shabble]

You could potentially store them inside/along with a virtual machine image which would be capable of rendering them, and then document the crap out of the VM (which would require as few, and as simple, external interfaces as possible), such that it could be easily re-created/ported to future OSes.

I think quite a few places where continuity over decade timescales (especially in the embedded world) are storing VM images with the exact versions of various parts of the toolchain they use, to ensure compatibility.

And of course, there's plenty of precedent with things like the Hercules IBM mainframe emulator[1], as well as things like VMWare, Bochs, QEMU, etc.

I think I've come across old DOS games being re-released wrapped with dosbox or something similar as a compat layer for modern OSes.

[1] https://en.wikipedia.org/wiki/Hercules_%28emulator%29