[Python-Dev] Decimal type question [Prothon]

["Martin v. Löwis"]

> Can anyone see any problem with this scheme?

How does that interact with (binary) floating point numbers?

I think the numerical people might dislike that approach, as
all their algorithms, C interfaces, and so on are heavily tied
to IEEE-754. Not being a NumPy user myself, I cannot say how
significant that is, and what could be done about it.

For example, would it be possible to automatically fall back
to binary floating point if the result cannot be represented
exactly (which would be most divide operations)? Would that
help?

Regards,
Martin