4 ms·

My brother the academic librarian / archivist has, for many years, been warning about this kind of obsolescence. It's necessary for we software folk to ask our

My brother the academic librarian / archivist has, for many years, been warning about this kind of obsolescence. It's necessary for we software folk to ask ourselves "two centuries from now, how will people make sense of our digital files?" One of the challenges is, of course, that it isn't obvious to any generation of creators what will be of lasting value. Academic archivists and MLS degree holders can't do it all, any more than they could control the 19th-century paper factories that churned out all that junk paper that turned brittle in a few decades.

thfuran 4y ago
Two centuries? I don't have an optical drive or zip drive or any flavor of floppy drive anymore. Lots of files from the last two decades are inaccessible to me even before considering whether they're still parseable / executable.
- antisthenes 4y ago
  I think we can safely view the 1980s/1990s as the testing grounds for finding out what works and what doesn't. Naturally, the things that worked were acquired and consolidated by larger and larger companies, which allowed tech to be cheap and commonplace. The #1 necessary factor for something being accessible later on is the sheer amount of popularity it enjoyed in the past. E.g. I think some common interfaces, like SATA / USB Type-A will never die, willing to take a $10k long bet that they will still be easily accessible in 15-20 years. Maybe not as popular as some newer ones like M.2 or USB-C. > Lots of files from the last two decades are inaccessible to me Yes, but not by volume. The amount of files ever stored on all 3.5 floppies in the world could very easily fit into a single server rack these days. Are these 0.01% of files more valuable in proportion than others? If not, then it makes 0 sense to focus on preserving those interfaces.
  - thfuran 4y ago
    Is the data produced in the past month more valuable than, say, all the writings of the 1st century just because there's more of it? Would the data produced in the first week of 2099 be worth more than all of history prior to 1900 just because there's more?
    - antisthenes 4y ago
      I think the point is that small data can be continuously carried over to new mediums if it's valuable enough. How many pre 1st century writings will fit on a 20TB hard drive? I'd venture to say - all of them. > Would the data produced in the first week of 2099 be worth more than all of history prior to 1900 just because there's more? No, and I think that's why we're mostly in agreement. Valuable and small data will always have accessible interfaces, because it is propagated throughout many mediums. If there's an apocalypse scenario looming, you bet your butt that someone will be printing the most valuable information to good old paper, in case all machine-readable interfaces vanish overnight (unlikely scenario that I'm not going to lose sleep over).
int_19h 4y ago
I ask this myself all the time. One thing that I think we need to adopt as standard practice is self-describing storage formats. That is, for any file (or, more broadly, a binary blob - say, hard drive image), I should be able to look for a well-known marker in the bitstream, and, having found it, be able to easily extract a detailed human-readable spec for the format with no special tools. Better yet if it comes with a simple reference implementation optimized for simplicity and ease of understanding, even if it's very slow. This doesn't solve the problem for physical representations of data, but so long as we can keep re-archiving the bits...

thfuran 4y ago

Two centuries? I don't have an optical drive or zip drive or any flavor of floppy drive anymore. Lots of files from the last two decades are inaccessible to me even before considering whether they're still parseable / executable.

antisthenes 4y ago
I think we can safely view the 1980s/1990s as the testing grounds for finding out what works and what doesn't. Naturally, the things that worked were acquired and consolidated by larger and larger companies, which allowed tech to be cheap and commonplace. The #1 necessary factor for something being accessible later on is the sheer amount of popularity it enjoyed in the past. E.g. I think some common interfaces, like SATA / USB Type-A will never die, willing to take a $10k long bet that they will still be easily accessible in 15-20 years. Maybe not as popular as some newer ones like M.2 or USB-C. > Lots of files from the last two decades are inaccessible to me Yes, but not by volume. The amount of files ever stored on all 3.5 floppies in the world could very easily fit into a single server rack these days. Are these 0.01% of files more valuable in proportion than others? If not, then it makes 0 sense to focus on preserving those interfaces.
- thfuran 4y ago
  Is the data produced in the past month more valuable than, say, all the writings of the 1st century just because there's more of it? Would the data produced in the first week of 2099 be worth more than all of history prior to 1900 just because there's more?
  - antisthenes 4y ago
    I think the point is that small data can be continuously carried over to new mediums if it's valuable enough. How many pre 1st century writings will fit on a 20TB hard drive? I'd venture to say - all of them. > Would the data produced in the first week of 2099 be worth more than all of history prior to 1900 just because there's more? No, and I think that's why we're mostly in agreement. Valuable and small data will always have accessible interfaces, because it is propagated throughout many mediums. If there's an apocalypse scenario looming, you bet your butt that someone will be printing the most valuable information to good old paper, in case all machine-readable interfaces vanish overnight (unlikely scenario that I'm not going to lose sleep over).

int_19h 4y ago

I ask this myself all the time. One thing that I think we need to adopt as standard practice is self-describing storage formats. That is, for any file (or, more broadly, a binary blob - say, hard drive image), I should be able to look for a well-known marker in the bitstream, and, having found it, be able to easily extract a detailed human-readable spec for the format with no special tools. Better yet if it comes with a simple reference implementation optimized for simplicity and ease of understanding, even if it's very slow. This doesn't solve the problem for physical representations of data, but so long as we can keep re-archiving the bits...