Yikes. If somebody hasn’t written a “falsehoods programmers believe about human writing systems” document this would make for a good start.
They had that, but needed a font that used a different number of strokes for the characters because of the superstition.
On further review, I think this is als similar to #12 & #13 on the list: "names are case-sensitive," and "names are not case-sensitive." To generalize that to include non-Western alphabets: display variations of the same character are significant, and display variations of the same character are not significant.
This of course goes back to the evergreen philosophical question "what even is a character, anyways?" Since we've found a case where two characters which are the same character are not the same character. Are they distinct characters or typographical variants? Yesn't: one would want them unified for searching, but distinct for printing.
But regardless of what they are, these characters/variants only show up in names. Names tend to retain archaic (or extinct) language variations longer than speech, which is the reason for rule #11, which is at least part of the problem.
I think these language examples are so good, as examples, because all aspects of them are clear and easy to follow. I think computerization of business and society and the systems that make them work, causes immense amounts of this kind of friction and pain all the time, in ways that are much harder to understand, explain, or catalog (which is precisely why it's such a big problem, though as far as I know it's received little attention)
[EDIT] To distill it, I think that trying to make a computer a "source of truth" rather than a tool, tends to do substantial violence to the "truth".
Language and writing exist to communicate, using patterns of signals that have shared meaning and recognition; things like alphabets and vocabularies are effectively (loose, overlapping, diasporic) consensus-state autoencoding models. They only work to compress meaning, when there are rules for said compression that generalize, and which don't have as many exceptions with their own separate symbols as there are words/names needing to be encoded.
Most countries don't allow you to just make up your own novel graphemes when writing a name on a birth certificate. And nobody is asking for that, either. (Presumably because living in a world where that was allowed would be horrible: you'd no longer being able to error-correct when reading, because any given mysterious squiggle in the middle of a word or name, might be exactly what some unknown-to-you-or-anyone-other-than-the-author character is supposed to look like. Is that "o with a curlicue" written here just a semi-cursive attempt at writing an "o" — or is it an "o" with a novel accent marker, one that appears nowhere else, but which must be preserved nevertheless to properly record this person's name?)
Instead, legal names are (in every country I'm aware of) required to be spelled using the character-set of the country you're entering a legal relationship with by being born / immigrating / etc. America? Legal names using the Latin alphabet. Japan? Legal names using characters from this set: https://en.wikipedia.org/wiki/Jinmeiy%C5%8D_kanji
Note, though, that legal names are representations of names. They aren't encodings of names. Your legal name is a distinct thing from your name, just as your credit-card number is a distinct thing from your name. It's an applied-for + registered + assigned systematic identifier for you — a bit like a domain name, or a vanity license-plate number. Which means that your legal name is not a lossy or lossless encoding of your name. It's, per se, a nickname. It doesn't have to have anything to do with your name. (And it often doesn't; immigrants often choose legal names entirely distinct from what they / their home country thinks of as their name.)
> do substantial violence to the "truth"
I don't think this kind of wild escalatory rhetoric is at all helpful to the (presumably) good cause intended. Probably the opposite actually.
Most of the time people are not in fact having substantial harm inflicted upon them by others out to do violence. That mindset seems incredibly fragile, paranoid, and divisive.
The reality is people are just working away to improve things and sometimes they make mistakes or don't have perfect information ahead of time, and other times making things better simply necessitates that not everybody's last whim can always be accommodated. The healthy mindset to have is that not every real or perceived slight against you is done because you are being persecuted by violent hatemongers, and that people should be more accommodating and accepting of the reality that systems and procedures designed for the benefit of everybody may just not be able to accommodate every unique request they have.
Webster's 1913:
> 2. Injury done to that which is entitled to respect, reverence, or observance; profanation; infringement; unjust force; outrage; assault.
> We can not, without offering violence to all records, divine and human, deny an universal deluge. - T. Burnet.
It's a bit more poetic a use, but it's not escalatory in the way you suggest.
[EDIT] Further, my point is (and I think that was clear?) that computerization necessarily does these things if you treat the computer as correct and humans as suspect, not that anyone's doing this on purpose.
"I'm going to shoot you" is escalatory. It doesn't matter that I could have been talking about photographing you.