Have been fighting this for years. All spreadsheets (including LO Calc) use the standard IEEE or hardware float representation for numbers, then round off for display to the numeric format specified. Works well, but results in occasional “your answer is approximately nnnnn.nnnn” and comparison failure issues. IBM mainframes (and some old Borland languages) had packed-decimal or BCD math that was exact for specified decimal precision and rounding - what a bookkeeper wants. I recently found a decimal data type in Python that does something similar. Is there a way to use that concept in LO Calc without wrapping the whole Python decimal library into an extension of some kind, and figuring out how to call it? NOT a Python expert so this is more a theoretical question than one I can jump into right now.
Understood: this would not be portable outside of LO and possibly OO. Might still be useful.