Floats better last?#1
Conversation
|
It's a rather trivial change, but I wasn't even sure why you would need so much to change Float64. But if you do, would you want to change Float32, and possibly Float16 too? I thought it might be awkward to have int in-between. Since the code is brand-new, and people haven't used much yet (and not registered) might be a good time to change, if you ever do. Could add keyword arguments too. |
|
Thanks for your suggestion! Actually, I'm not even sure I want to keep this positional-arguments method, the pairs version is clearer. The reasoning for how it is now is that I assumed In any case, I don't want to overload this method (which should remain simple) with
Actually this was first implemented with keyword arguments, and this is still available but not documented. But @cmcaine suggested that using |
|
After our discussion on Discourse, I would (at least for now) simply change floats to be unchanged as a default. It could be related to the PR change here, or not. It's still "safe", just surprising to beginners. At least BigFloats doesn't makes safer, just avoids overflow (not a big problem with floats), and has problems. |
ff4cd06 to
f8d8c22
Compare
No description provided.