Uhm, I haven’t programmed in a low level language in years. I use python for my job now, and all I know are floats and ints. I don’t know what this foreign language is you speak of.
I have been thinking that maybe modern programming languages should move away from supporting IEEE 754 all within one data type.
Like, we’ve figured out that having a null value for everything always is a terrible idea. Instead, we’ve started encoding potential absence into our type system with Option or Result types, which also encourages dealing with such absence at the edges of our program, where it should be done.
Well, NaN is null all over again. Instead, we could make the division operator an associated function which returns a Result<f64> and disallow f64 from ever being NaN.
My main concern is interop with the outside world. So, I guess, there would still need to be a IEEE 754 compliant data type. But we could call it ieee_754_f64 to really get on the nerves of anyone wanting to use it when it’s not strictly necessary.
Well, and my secondary concern, which is that AI models would still want to just calculate with tons of floats, without error-handling at every intermediate step, even if it sometimes means that the end result is a shitty vector of NaNs, that would be supported with that, too.
Nan isn’t like null at all. It doesn’t mean there isn’t anything. It means the result of the operation is not a number that can be represented.
The only option is that operations that would result in nan are errors. Which doesn’t seem like a great solution.
Float processing is at the hardware level. It needs a way to signal when an unrepresented value would be returned.
Well, that is what I meant. That NaN is effectively an error state. It’s only like null in that any float can be in this error state, because you can’t rule out this error state via the type system.
Why do you feel like it’s not a great solution to make NaN an explicit error?
Theres plenty of cases where I would like to do some large calculation that can potentially give a NaN at many intermediate steps. I prefer to check for the NaN at the end of the calculation, rather than have a bunch of checks in every intermediate step.
How I handle the failed calculation is rarely dependent on which intermediate step gave a NaN.
This feels like people want to take away a tool that makes development in the engineering world a whole lot easier because “null bad”, or because they can’t see the use of multiplying 1e27 with 1e-30.
While I get your proposal, I’d think this would make dealing with float hell. Do you really want to .unwrap() every time you deal with it? Surely not.
One thing that would be great, is that the / operator could work between Result and f64, as well as between Result and Result. Would be like doing a .map(|left| left / right) operation.
Well, not every time. Only if I do a division or get an ieee_754_f64 from the outside world. That doesn’t happen terribly often in the applications I’ve worked on.
And if it does go wrong, I do want it to explode right then and there. Worst case would be, if it writes random NaNs into some database and no one knows where they came from.
As for your suggestion with the slash accepting Results, yeah, that could resolve some pain, but I’ve rarely seen multiple divisions being necessary back-to-back and I don’t want people passing around a Result<f64> in the codebase. Then you can’t see where it went wrong anymore either.
So, personally, I wouldn’t put that division operator into the stdlib, but having it available as a library, if someone needs it, would be cool, yeah.
I agree with moving away from floats but I have a far simpler proposal… just use a struct of two integers - a value and an offset. If you want to make it an IEEE standard where the offset is a four bit signed value and the value is just a 28 or 60 bit regular old integer then sure - but I can count the number of times I used floats on one hand and I can count the number of times I wouldn’t have been better off just using two integers on -0 hands.
Floats specifically solve the issue of how to store a ln absurdly large range of values in an extremely modest amount of space - that’s not a problem we need to generalize a solution for. In most cases having values up to the million magnitude with three decimals of precision is good enough. Generally speaking when you do float arithmetic your numbers will be with an order of magnitude or two… most people aren’t adding the length of the universe in seconds to the width of an atom in meters… and if they are floats don’t work anyways.
I think the concept of having a fractionally defined value with a magnitude offset was just deeply flawed from the get-go - we need some way to deal with decimal values on computers but expressing those values as fractions is needlessly imprecise.
Based and precision pilled.
Float is bloat!
As a programmer who grew up without a FPU (Archimedes/Acorn), I have never liked float. But I thought this war had been lost a long time ago. Floats are everywhere. I’ve not done graphics for a bit, but I never saw a graphics card that took any form of fixed point. All geometry you load in is in floats. The shaders all work in floats.
Briefly ARM MCU work was non-float, but loads of those have float support now.
I mean you can tell good low level programmers because of how they feel about floats. But the battle does seam lost. There is lots of bit of technology that has taken turns I don’t like. Sometimes the market/bazaar has spoken and it’s wrong, but you still have to grudgingly go with it or everything is too difficult.
Floats make a lot of math way simpler, especially for audio, but then you run into the occasional NaN error.
all work in floats
We even have float16 / float8 now for low-accuracy hi-throughput work.
Even float4. You get +/- 0, 0.5, 1, 1.5, 2, 3, Inf, and two values for NaN.
Come to think of it, the idea of -NaN tickles me a bit. “It’s not a number, but it’s a negative not a number”.
I think you got that wrong, you got +Inf, -Inf and two NaNs, but they’re both just NaN. As you wrote signed NaN makes no sense, though technically speaking they still have a sign bit.
I’d have to boulder check, but I think old handheld consoles like the Gameboy or the DS use fixed-point.
IMO, floats model real observations.
And since there is no precision in nature, there shouldn’t be precision in floats either.
So their odd behavior is actually entirely justified. This is why I can accept them.
I just gave up fighting. There is no system that is going to both fast and infinitely precision.
So long ago I worked in a game middleware company. One of the most common problems was skinning in local space vs global space. We kept having customers try and have global skinning and massive worlds, then upset by geometry distortion when miles away from the origin.
How do y’all solve that, out of curiosity?
I’m a hobbyist game dev and when I was playing with large map generation I ended up breaking the world into a hierarchy of map sections. Tiles in a chunk were locally mapped using floats within comfortable boundaries. But when addressing portions of the map, my global coordinates included the chunk coords as an extra pair.
So an object’s location in the 2D world map might be ((122, 45), (12.522, 66.992)), where the first elements are the map chunk location and the last two are the precise “offset” coordinates within that chunk.
It wasn’t the most elegant to work with, but I was still able to generate an essentially limitless map without floating point errors poking holes in my tiling.
I’ve always been curious how that gets done in real game dev though. if you don’t mind sharing, I’d love to learn!
