I had a bunch of old Word-documents from a late researcher using non-latin scripts and custom fonts that needed fixing for archiving (the only important step wa

I had a bunch of old Word-documents from a late researcher using non-latin scripts and custom fonts that needed fixing for archiving (the only important step was to make these readable by humans). Don't remember what encoding. Current version of Word wouldn't even touch these, raising errors left and right. In the end I changed the strings that specified the problematic/missing fonts in a hex editor (I naively assumed that would do the least damage...) to a compatible font I eventually found after some digital archaeology. LibreOffice accepted these files but didn't display them correctly (as in: according to what my eyes and brain expected - "machine readable" was in a different universe at this point). I saved the hex-edited documents in whatever version of docx LibreOffice exports to. Word accepted these, displayed them as well as could be hoped, and a little bit of modern history could be saved for future generations. Not a very technical approach, I guess. I feel the 1990's in general did a lot of damage. It's not only Word. Competing, proprietary formats together with lacking text-encodings (and in my case custom fonts messing up character tables, due to academic/linguistic needs) has caused more information loss than the decades before and after for the data I've worked with. Suffice to say I use plain text as much as possible (markdown, latex), as text-encoding alone is hard enough.