[pmarreck]

> Text is no more robust to random bit flips

Erm, a single bitflip will ruin, at maximum, a single character, which probably has enough context around it to be easily manually repaired.

Not so pretty much any compressed, encrypted or binary format... either a large chunk of it gets undecodable or the whole thing is lost.

[something98]

Not universally true. But, regardless, most files are not text.

[OtomotO]

> But, regardless, most files are not text

Hence the initial post. (Are we running in circles now?)

[something98]

I'm pointing out that everyone is comfortable with most files not being human-readable. And that it's a red herring to suggest that text is the longest-lived format. I'm sure my data could outlive text files if I "stored" it on granite blocks, but it's not very useful in that format. Same with plain text; it just isn't a very useful format.

[lytex]

I think this web is quite revealing:

https://plain-text.co/#the-office-model-and-the-engineering-... https://plain-text.co/#two-revolutions-in-computing

It boils down to a tradeoff of learning your tools while retaining full control vs use apps that abstract away complexity by giving up control.

If your work needs some strict technical requisite (reproducibility for example, or any other of the -ilities in software engineering), you need to have the control to tweak your tools so that you ensure it (Engineering Model, plain text files), but mostly is not needed so you can get away with working in the Office Model (binary files)