Hurray for plain text. I also use troff (e.g., for my CV), and I expect that even if maintenance of groff stops tomorrow, I’ll still be able to use it on some variant of Unix for quite a while – after all, we can still build most Unix C code from the 1980s with little changes. And besides, like the author says, being a plain text based format, it’s readable as it is.
This is partly why I have switched to Orgmode for almost everything, including invoices.
Too, you can export from org mode through LaTeX to get really handsome documents, for those times when typography and layout matter (and they do, in many contexts.)
Sounds more like “use open formats when possible”.
I wouldn’t be confident of being able to get OpenOffice working 2 or 3 decades from now.
If you use an xml/latex/rtf/json/markdown document format then worst case you can at least get the text data and maybe some basic bold/italic/underline information out of it by writing a parser yourself, best case some application/library still understands it.
human-readable open formats - if all else fails you should be able to extract the content and/or write a renderer with just the file to look at
Plain doesn’t necessarily mean ASCII. Though I’d be wary of using combining family member characters and other funky features of Unicode in texts that I want to remain readable in 2100.
Eh, ASCII was made for teletypes and EBCDIC for punched cards and we can still read them nearly 80 years on. Even the funky characters like ENQ and SO/SI.
Unicode is a beast but its a well specified beast. Our progeny won’t have trouble with it.