Hacker News Reader

Comment by qwerty456127 | Hacker News Reader

I wonder what if someone could reproduce this kind of ancient high-tech using modern precision production technologies and durable materials. Could we build those huge size low density hard drives and other components with orders of magnitude higher quality so they would last forever and never need anything replaced?

View on HN

http://civboot.org is basically trying to answer exactly this question

Wow thanks! I always felt anxious about our civilization apparently loosing ways to build things from scratch. Nice to know I'm not alone and somebody is working on this problem.

Define forever. Especially because you mentioned hard drives - moving parts. Which will definitely wear out in time.

I have HDDs (spining rust) in my desktop that are 10+ years and still running fine. I had older SCSI UW-320 disks that were running 20+ years until well, started to give I/O errors. Workloads are low here (mostly read). So I imagine it is doable to do HDD with low density and high grade components that would last 100 years. Interesting challenge imo.

I suppose the modern equivalent of what GP is asking for would be SLC flash. Each bit has its own dedicated cell, compared to the denser and more common MLC/TLC/QLC flash where cells are shared. I would trust data stored in SLC to last unpowered for a lot longer than your average QLC disk.

Or maybe core memory would be more stable over time?

If we're going for ultimate durability, maybe the laser inscribed glass would be the way to go, but not sure it's practical as typical computer memory. [0]

[0] https://www.sciencealert.com/this-new-5d-data-storage-disc-c...

As far as I know hard drives are more and more prone to have bad blocks as you travel back in time. An old man who had access to computers before I have even been born told me every HDD in production had some. However, nowadays as HDDs density went insane and each block is so tiny, a single bad block usually means a red alert suggesting more are coming and the whole drive is dying, nevertheless this is a rare thing to happen as we now can produce all the parts with extreme precision making it a "swiss clockwork" sort of reliable. This way, can't we produce a flawlessly perfect HDD of the old kind so its huge blocks would never turn bad?

I don't think modern hdd bad blocks work this way.

Look at the SMART property "reallocated sector count", which was there before SSDs.

They have firmware and spare capacity just like SSDs. When you do get a bad block it means they ran out of spare space.

Reply on news.ycombinator.com