Hi everyone!

[NotFitForPurpose]

Hi,
this is very much from "memory" for devices circa that ERA:
FD's ~ 3 to 5 years (Magnetic)
HD's ~ 3 to 5 years (Magnetic)
Tape ~ 3 to 5 years (Magnetic)
CD's ~ 10 to 20 years (Optical)
EPROM's - 10 to 20 years (Charge)
ROM (Fusible link) ~ 50+ years (Physical Link)
Core Storage ~ I can't remember (magnetic)
Paper tape ~ 15 to 30 (physical holes)
SSD's - 5 to 20 years (Charge)
USB Stick's (EEPROM) ~ 10 to 15 years (Charge)

There are a few others, I think these are the most common used around the 1980's / 1990's.
When I was designing systems, Fusible link ROM's and to some extent Mask programmed ROM's were the gold standard.

All theoretical and your milage will of course vary. Certainly I have floppy disks and hard disks that have lasted much longer. Though I wouldn't have bet my future pension on it!

For the floppy disks, the drive is an important part of the process. If you can try another drive, you might have some success.

Taking an EPROM (the typical Data Packs), I suspect it more likely the number and quality of the erase (formatting) and write cycles will determine if you will loose data than the device its self.

[Cosi]

I find the idea of UV formatting really quite charming (...)

It's surprisingly "analog" – more like baking bread or smelting rather than something you'd typically associate with computers

[Caitlin]

Hi,
this is very much from "memory" for devices circa that ERA:
FD's ~ 3 to 5 years (Magnetic)
HD's ~ 3 to 5 years (Magnetic)
Tape ~ 3 to 5 years (Magnetic)
CD's ~ 10 to 20 years (Optical)
EPROM's - 10 to 20 years (Charge)
ROM (Fusible link) ~ 50+ years (Physical Link)
Core Storage ~ I can't remember (magnetic)
Paper tape ~ 15 to 30 (physical holes)
SSD's - 5 to 20 years (Charge)
USB Stick's (EEPROM) ~ 10 to 15 years (Charge)

There are a few others, I think these are the most common used around the 1980's / 1990's.
When I was designing systems, Fusible link ROM's and to some extent Mask programmed ROM's were the gold standard.
<snip>

Very interesting! The hdds from the early-mid 90s I have are not happy at all, though the CDs seem to be keeping on fine (not so for my CD-Rs from the 2000s, which have become unreadable in many cases). Paper tape is interesting -- is that due to deterioration of the paper stock?

[NotFitForPurpose]

Hi Catilin,
I am sure there will be more detail on the web.
Here's a summary, Silicon semiconductors are sensitive to light. Possibly most recently 'highlighted' in the press by someone taking a photo of a Raspberry Pi that reset as a result of the camera flash shutting down a voltage regulator.

"likely due to the U16 chip, a voltage regulator, being sensitive to light. This light sensitivity can cause the Pi to shut down or reset. Additionally, a Raspberry Pi 2 is particularly susceptible to this issue.".

Again, when I first started. To get a light sensor, scratch the paint off the glass cover of a transistor.

Back to EPROM's, so U.V. light has enough energy to allow current to leak from the capacitor in the memory cell that stores charge for the data bits to the substrate. "The EPROM is exposed to ultraviolet (UV) light, which removes the trapped electrons and resets the cells to '0'""

The smell of Ozone should be relatively small when 'baking' an EPROM. It helps a lot not to over or under 'cook' / expose them.
Under exposure leads to the possibility of random bit's, over exposing can degrade the device so they loose the capability to store them robustly (having taken great care to never having done it, I can't speak personally to that).

Your CD's may not work if they are certain types, as the driver has probably been deprecated. I managed to recover an old HP CD R/W disk with a heap of work using a Samsung Writemaster and custom driver to get at the ISO 9660 format CD's. All my old PSION stuff was on it. It took some digging but I manged to get free read only mode software. It was easier to boot the Windows XP machine the old drive was installed in though.

For punch tape its likely associated with poor storage and tape stretch due to use. I did see one article in the past claiming 1000 years for punch tape; perhaps they were taking it to a warm dry country and burying it under a big stone built pyramid .

[thesourcerer]

It is interesting that EPROMs have such a low failure rate. I think it is less than 0.1%, having had over 10,000 now, all over 30 years old, and only a handful of failures. Considering that some may have been overexposed or badly stored or handled etc it shows their inherent longevity.
I seem to remember the estimated life span was 100 years back in the late 80’s?
Time will tell, but I won’t be around to confirm that!

[Caitlin]

That's an impressively low failure rate! 100 years would be impressive -- maybe there is hope for digital archiving after all....! Punch tape at 1000 years is even more impressive, but yes, going to need to be in very specific circumstances - I also wonder if that applies to 'modern' wood-pulp paper (cloth-based paper, the historic type, has a much longer predicted life than modern wood-pulp paper, iirc).

The points on making sure that they get the right amount of UV light and not too much nor too little is interesting; I guess that's why one doesn't just stick them on the window sill for a week?!

My worse CD-Rs, made in about 2007-09, have seen the dye-layer (?) change colour and degrade, rendering them unreadable...