The two situations are not at all comparable. "Making parentheses mandatory" for `print` was a consequence of making it a function rather than an entire separate statement. This was a radical syntax change with far-reaching implications:

* The special syntax for redirecting to a file could no longer work, so that became a keyword argument

* Using a trailing comma to suppress a trailing newline could no longer work, so that became a keyword argument too

* A keyword argument could also be added to control spacing between items - there was no reasonable way to express that in the old syntax (and the old syntax used some complex heuristics here to try to "do the right thing")

* Being a function, `print` can now be used as a higher-order function (I have done this)

* Being a function, arguments to `print` can be taken from a sequence, even e.g. a generator, using the `*` syntax (I have done this)

It would have been completely impossible to preserve the old semantics while ending up with a function. Offering both new and old semantics would have required using a different name for the new function, creating parallel but subtly incompatible ways of doing the same thing and completely defeating the purpose.

The proposal here, on the other hand, is simply to deprecate (implying eventual removal) a method (`__mod__`) from the built-in string type. That would break existing code and create churn, but it doesn't appear to offer anything on the other side of the equation except for a tiny reduction in the size of the CPython code base. (It seems rather unlikely to me that the bit of code in question sees frequent changes, so it isn't creating a maintenance burden. It's just inelegant - and while I personally value elegance rather highly, I would still draw a line here.)

Well, it would offer the OP the ability to use the operator for something else. But I strain to imagine the use case OP is describing in that thread. There's a long tradition of understanding the `%` operator this way (thanks to C's decision to use the % symbol for formatting codes in `printf`), so it would be confusing.

More importantly, though, IMO: OP describes a math library, and in ordinary mathematical interpretations, the modulo operator is not commutative. With two operands of the same type, x % y is not expected to equal y % x; with different types, I don't see why it's expected that both are even possible.