Actually, it may not be as out of line if you think a bit differently. If you can show, or reasonably believe that Float64 representation is insufficient (and there are numerous spots where that can be true), then it isn't illogical to think of stepping up to a 128-bit representation (since they do tend to come in powers of two), If the domain is Financial, (or something else that deals with exact decimal fractions) then thinking of at the same time switch to a decimal representation makes sense. Decimal64 MIGHT work, but world-scale numbers start to come very close to running out of precision (you only get 16 digits), especially if doing calculations that need fractions of pennies for things like interest calculation. And Decimal64 isn't that much of a help, as SQLite doesn't support it either (at least out of the box). You don't need the full resolution of the new format to be reasonable to need, you just need to show that the next option down is possibly insufficient. Maybe a less standard format more akin to Float/Decimal-80/96 might make sense, but then are you backing yourself into a corner with something not as standard.