This got me curious, and yeah it turns out Elm's dictionary implementation uses values, not pointers when retrieving values.
elm repl
---- Elm 0.19.1 ----------------------------------------------------------------
Say :help for help and :exit to exit! More at <https://elm-lang.org/0.19.1/repl>
--------------------------------------------------------------------------------
> import Dict exposing (Dict)
> nan = 0/0
NaN : Float
> nan
NaN : Float
> nan == nan
False : Bool
> naan = 0/0
NaN : Float
> d = Dict.fromList [(nan, "a"), (naan, "b")]
Dict.fromList [(NaN,"a"),(NaN,"b")]
: Dict Float String
> Dict.toList d
[(NaN,"a"),(NaN,"b")]
: List ( Float, String )
> Dict.keys d
[NaN,NaN] : List Float
> Dict.get nan d
Nothing : Maybe String
> Dict.get naan d
Nothing : Maybe String
Yeah IEEE 754 floating point numbers should probably not be hashable, and the weird (but standard-defined) behaviour with respect to NaN equality is one good reason for this.
Nor is that inequality an oddity at all. If you were to think NaN should equal NaN, that thought would probably stem from the belief that NaN is a singular entity which is a misunderstanding of its purpose. NaN rather signifies a specific number that is not representable as a floating point. Two specific numbers that cannot be represented are not necessarily equal because they may have resulted from different calculations!
I'll add that, if I recall correctly, in R, the statement NaN == NaN evaluates to NA which basicall means "it is not known whether these numbers equal each other" which is a more reasonable result than False.
It's the only "primitive type" that does that. If I deserialize data from wire, I'll be very surprised when the same bits deserialize as unequal variables. If it cannot be represented, then throwing makes more sense than trying to represent it.
Other primitive types also do this, but this is not clearly visible from high-level programming languages, because most HLLs have only incomplete support for the CPU hardware.
If you do a (signed) integer operation, the hardware does not fit the result in a register of the size expected in a HLL, but the result has some bits elsewhere, typically in a "flags" register.
So the result of an integer arithmetic operation has an extra bit, usually named as the "overflow" bit. That bit is used to encode a not-a-number value, i.e. if the overflow bit is set, the result of the operation is an integer NaN.
For correct results, one should check whether the result is a NaN, which is called checking for integer overflow (unlike for FP, the integer execution units do not distinguish between true overflow and undefined operations). After checking that the result is not a NaN, the extra bit can be stripped from the result.
If you serialize an integer number for sending it elsewhere, that implicitly assumes that wherever your number was produced, someone has tested for overflow, i.e. that the value is not a NaN, so the extra bit was correctly stripped from the value. If nobody has tested, your serialized value can be bogus, the same as when serializing a FP NaN and not checking later that it is a NaN, before using one of the 6 relational operators intended for total orders, which may be wrong for partial orders.
> "it is not known whether these numbers equal each other"
Equality, among other operations, are not defined for these inputs. NaN's really are a separate type of object embedded inside another objects value space. So you get the rare programmers gift of being able to construct a statement that is not always realizable based solely on the values of your inputs.
It is not an IEEE 754 oddity. It is the correct mathematical behavior.
When you define an order relation on a set, the order may be either a total order or a partial order.
In a totally ordered set, there are 3 values for a comparison operation: equal, less and greater. In a partially ordered set, there are 4 values for a comparison operation: equal, less, greater and unordered.
For a totally ordered set you can define 6 relational operators (6 = 2^3 - 2, where you subtract 2 for the always false and always true predicates), while for a partially ordered set you can define 14 relational operators (14 = 2^4 - 2).
For some weird reason, many programmers have not been taught properly about partially-ordered sets and also most programming languages do not define the 14 relational operators needed for partially ordered sets, but only the 6 relational operators that are sufficient for a totally ordered set.
It is easy to write all 14 relational operators by combinations of the symbols for not, less, greater and equal, so parsing this in a programming language would be easy.
This lack of awareness about partial order relations and the lack of support in most programming languages is very bad, because practical applications need very frequently partial orders instead of total orders.
For the floating-point numbers, the IEEE standard specifies 2 choices. You can either use them as a totally-ordered set, or as a partially-ordered set.
When you encounter NaNs as a programmer, that is because you have made the choice to have partially-ordered FP numbers, so you are not allowed to complain that this is an odd behavior, when you have chosen it. Most programmers do not make this choice consciously, because they just use the default configuration of the standard library, but it is still their fault if the default does not do what they like, but nonetheless they have not changed the default settings.
If you do not want NaNs, you must not mask the invalid operation exception. This is actually what the IEEE standard recommends as the default behavior, but lazy programmers do not want to handle exceptions, so most libraries choose to mask all exceptions in their default configurations.
When invalid operations generate exceptions, there are no NaNs and the FP numbers are totally ordered, so the 6 relational operators behave as naive programmers expect them to behave.
If you do not want to handle the invalid operation exception and you mask it, there is no other option for the CPU than to use a special value that reports an invalid operation, and which is indeed not-a-number. With not-numbers added to the set of FP numbers, the set becomes a partially-ordered set and all relational operators must be interpreted accordingly.
If you use something like C/C++, with only 6 relational operators, then you must do before any comparison tests to detect any NaN operand, because otherwise the relational operators do not do what you expect them to do.
In a language with 14 relational operators, you do not need to check for NaNs, but you must choose carefully the relational operator, because for a partially-ordered set, for example not-less is not the same with greater-or-equal (because not-less is the same with greater-or-equal-or-unordered).
If you do not expect to do invalid operations frequently, it may be simpler to unmask the exception, so that you will never have to do any test for NaN detection.
IEEE 754 prescripes, for better or worse, that any mathematical comparison operator (==, <, > ….) involving at least one NaN must always return false, including comparison against itself. This is annoying for something like dictionaries or hashtables. C# has a solution: if you call a.Equals(b) on two floats a and b, it will return true also if both are NaN. I think this is a cool solution: it keeps the meaning of math operators the same identical with other languages, but you still have sensible behavior for containers. I believe this behavior is copied from Java.
I guess because the hash of an instance stays consistent (which is used to retrieve the value from the dict). The `__eq__` method must disregard the hash and return False for all nans.
But the hash alone shouldn't be enough to match the key. Isn't an equality check also needed to avoid a false positive? That's the idea behind a hash table, as I understand it. (I'm not a Python programmer.)
> Last week in the Python Discord we had an unusual discussion about a Python oddity.
Oh, I missed it. But yes, this is more to do with NaN than Python.
> But, of course, you can't actually get to those values by their keys: ... That is, unless you stored the specific instance of nan as a variable:
Worth noting that sets work the same way here, although this was glossed over: you can store multiple NaNs in a set, but not the same NaN instance multiple times. Even though it isn't equal to itself, the insertion process (for both dicts and sets) will consider object identity before object equality:
>>> x = float('nan')
>>> {x for _ in range(10)}
{nan}
And, yes, the same is of course true of `collections.Counter`.
If I could change one thing in computing, it'd be how SQL handles NULL. But if I got a second thing, it'd be how IEEE handles NaN. I probably wouldn't even allow NaN as a representation. If some mathematical operation results in what would be NaN, I'd rather force the programming language to throw some sort of interrupt or exception. Much like what happens when you divide an integer by 0. Heck, I'd probably even stop infinity from being represented with floats. If someone did 1/0 or 0/0, I'd interrupt rather than generating an INF or NaN.
In my experience, INF and NaN are almost always an indicator of programming error.
If someone want's to programmatically represent those concepts, they could do it on top of and to the side of the floating point specification, not inside it.
This is also why Rust has separate PartialEq and Eq traits - the latter is only available for types that don't have weird not-self-equal values like floating point NaNs or SQL NULLs. If you lie to Rust and create a wrapper type over f32 or f64 that has Eq, then you'd get unindexable NaN keys that just sit in your hashmap forever.
The real surprise to me is that Python can index NaN keys sometimes, at least by reference to the original NaN. I knew CPython does some Weird Shit with primitive values, so I assume it's because the hashmap is comparing by reference first and then by value.
`If an operation has a single NaN input and propagates it to the output, the result NaN's payload should be that of the input NaN (this is not always possible for binary formats when the signaling/quiet state is encoded). If there are multiple NaN inputs, the result NaN's payload should be from one of the input NaNs; the standard does not specify which.'
maybe it's that multiple bit patterns can be NaN and these are two different ones? In IEEE-754, a number with all the exponent bits set to 1 is +/-infinity if the fraction bits are all zero, otherwise it's NaN. So these could be values where the fractions differ. Can you see what the actual bits it's setting are?
If NaNs were meant to represent unknown quantities, then they would return false for all comparisons. But NaN != NaN is true. Assuming that two unknowns are always different is just as incorrect as assuming that they're always the same.
I'd also push back on the idea that this behavior makes sense. In my experience it's a consistent source of confusion for anyone learning to program. It's one of the clearest violations of the principle of least astonishment in programming language design.
As others have noted, it makes conscientious languages like Rust do all sorts of gymnastics to accommodate. It's a weird edge case, and imo a design mistake. "Special cases aren't special enough to break the rules."
Also, I think high level languages should avoid exposing programmers to NaN whenever possible. Python gets this right: 0/0 should be an error, not a NaN.
As a standard for floating point representation and computation, IEEE 754 solved multiple long-standing serious problems better than anything that came before it. I don't think its sensible to judge it with a PL design lens like "principle of least astonishment"; certainly not as if IEEE754 is a peer to Rust or Python. Or, you could learn about the surprise, frustration, and expense involved in trying to write portable numeric code in the 1970s, prior to IEEE 754:
I like this justifiation of NaN != NaN; it emphasizes that NaN has representional intent, more than just some bit pattern.
We take for granted that (except for things like x86 extended precision registers) floating point basically works the same everywhere, which was the huge victory of IEEE 754. It easy to lose sight of that huge win, and to be ungrateful, when one's first introduction to IEEE 754 are details like NaN!=NaN.
Python is extra annoying though with refusing to support division through zero the way other programming languages with IEEE floats do (i.e. output inf or nan instead of throwing an exception), even though it has no problem doing things like float('inf') / float('inf') --> nan. It specifically does it for division through zero as if it wants to be a junior grade calculator just for this one thing. They could at least have fixed this when breaking backwards incompatibility from python2 to 3...
Nah. Python gets it right; all high level languages should operate this way. Division by zero is a bug 90% of the time. Errors should never pass silently. Special cases aren't special enough to break the rules.
IEEE floats should be a base on which more reasonable math semantics are built. Saying that Python should return NaN or inf instead of throwing an error is like saying that Python should return a random value from memory or segfault when reading an out-of-bounds list index.
In most languages, `x: float = 0` involves an implicit conversion from int to float. In Python, type annotations have no impact on runtime behavior, so even though the type checker accepts this code, `type(x)` will be `int` -- python acts as if `int` was a subtype of `float`.
It would be weird if the behavior of `1 / x` was different depending on whether `0` or `0.0` was passed to a `x: float` parameter -- if `int` is a subtype of `float`, then any operation allowed on `float` (e.g. division) should have the same behavior on both types.
This means Python had to choose at least one:
1. division violates the liskov substitution principle
2. division by zero involving only integer inputs returns NaN
3. division by zero involving only float inputs throws
exception
4. It's a type error to pass an int where a float is expected.
They went with option 3, and I think I agree that this is the least harmful/surprising choice. Proper statically typed languages don't have to make this unfortunate tradeoff.
C does different things for 0.0 / 0.0 and 0 / 0 and it's not that weird to deal with (well it has other issues like it being platform dependent what happens with this). JS has no problem with it either (0.0 / 0.0 gives nan, 0n / 0n gives exception since it are integers).
Python is the only language doing this (of the ones I use at least).
I don't think the notation `x: float = 0` existed when it was new by the way so that can't be the design reason?
since python seems to handle integer through integer divisions as float (e.g. 5 / 2 outputs 2.5), 0 / 0 giving nan would seem to be expected there
> liskov substitution principle
that would imply one is a subtype of another, is that really the case here? there are floats that can't be represented as an integer (e.g. 0.5) and integers that can't be represented as a double precision float (e.g. 18446744073709551615)
Python chose, quite some time ago, not to follow C's lead on division:
PEP 238 – Changing the Division Operator (2001) [1]
The rationale is basically that newcomers to Python should see the results that they would expect from grade school mathematics, not the results that an experienced programmer would expect from knowing C. While the PEP above doesn't touch on division by zero, it does point toward the objective being a cohesive, layman-friendly numeric system.
There's no non-confusing option for comparisons. You have two invalid values, but they aren't necessarily the same invalid value. There are multiple operations that can produce NaN.
It's a sentinel value for an error. Once you have an error, doing math with the error code isn't sensible.
There are no non-confusing options, but some of those are still clearly worse than others.
What should sorted([3, nan, 2, 4, 1]) give you in Python?
A) [1, 2, 3, 4, nan] is an good option
B) [nan, 1, 2, 3, 4] is an good option
C) An error is an good option
D) [3, nan, 1, 2, 4] is a silly, bad option. It's definitely not what you want, and it's quiet enough to slip by unnoticed. This is what you get when Nan != NaN
NaN == NaN is wrong. NaN != NaN is wrong, unintuitive, and breaks the rest of your code. If you want to signal that an operation is invalid, then throw an error. The silently nonsensical semantics of NaN are the worst possible response
Actually none of the floating point value are normal :D
Although Almost All† Real Numbers are Normal, none of the floats are normal, and nor are most numbers any regular person would think of (possible exceptions Pi and Euler's Number which are conjectured to be Normal although it is unproven)
† Almost All is a term of art in mathematics, remember there are uncountably many real numbers, so the fact the Normals are also uncountable puts them at least in the right ballpark, unlike the rationals.
This got me curious, and yeah it turns out Elm's dictionary implementation uses values, not pointers when retrieving values.
> we had an unusual discussion about a Python oddity
There are so many discussions about "X language is so weird about it handles numbers!" and it's just IEEE 754 floats.
The oddity here is not the float itself, it's that Python provided a default hash implementation for floats
Yeah IEEE 754 floating point numbers should probably not be hashable, and the weird (but standard-defined) behaviour with respect to NaN equality is one good reason for this.
Another fun one is -0.0 == 0.0 but they have different bit patterns, though python seems to handle this case and they both hash to 0.
NaN that is not equal to itself _even if it's the same variable_ is not a Python oddity, it's an IEEE 754 oddity.
It's an IEEE-754 oddity that Python chose to adopt for its equality.
IEEE-754 does remainder(5, 3) = -1, whereas Python does 5 % 3 = 2.
There's no reason to expect exact equivalence between operators.
Nor is that inequality an oddity at all. If you were to think NaN should equal NaN, that thought would probably stem from the belief that NaN is a singular entity which is a misunderstanding of its purpose. NaN rather signifies a specific number that is not representable as a floating point. Two specific numbers that cannot be represented are not necessarily equal because they may have resulted from different calculations!
I'll add that, if I recall correctly, in R, the statement NaN == NaN evaluates to NA which basicall means "it is not known whether these numbers equal each other" which is a more reasonable result than False.
It's the only "primitive type" that does that. If I deserialize data from wire, I'll be very surprised when the same bits deserialize as unequal variables. If it cannot be represented, then throwing makes more sense than trying to represent it.
Other primitive types also do this, but this is not clearly visible from high-level programming languages, because most HLLs have only incomplete support for the CPU hardware.
If you do a (signed) integer operation, the hardware does not fit the result in a register of the size expected in a HLL, but the result has some bits elsewhere, typically in a "flags" register.
So the result of an integer arithmetic operation has an extra bit, usually named as the "overflow" bit. That bit is used to encode a not-a-number value, i.e. if the overflow bit is set, the result of the operation is an integer NaN.
For correct results, one should check whether the result is a NaN, which is called checking for integer overflow (unlike for FP, the integer execution units do not distinguish between true overflow and undefined operations). After checking that the result is not a NaN, the extra bit can be stripped from the result.
If you serialize an integer number for sending it elsewhere, that implicitly assumes that wherever your number was produced, someone has tested for overflow, i.e. that the value is not a NaN, so the extra bit was correctly stripped from the value. If nobody has tested, your serialized value can be bogus, the same as when serializing a FP NaN and not checking later that it is a NaN, before using one of the 6 relational operators intended for total orders, which may be wrong for partial orders.
> "it is not known whether these numbers equal each other"
Equality, among other operations, are not defined for these inputs. NaN's really are a separate type of object embedded inside another objects value space. So you get the rare programmers gift of being able to construct a statement that is not always realizable based solely on the values of your inputs.
It is not an IEEE 754 oddity. It is the correct mathematical behavior.
When you define an order relation on a set, the order may be either a total order or a partial order.
In a totally ordered set, there are 3 values for a comparison operation: equal, less and greater. In a partially ordered set, there are 4 values for a comparison operation: equal, less, greater and unordered.
For a totally ordered set you can define 6 relational operators (6 = 2^3 - 2, where you subtract 2 for the always false and always true predicates), while for a partially ordered set you can define 14 relational operators (14 = 2^4 - 2).
For some weird reason, many programmers have not been taught properly about partially-ordered sets and also most programming languages do not define the 14 relational operators needed for partially ordered sets, but only the 6 relational operators that are sufficient for a totally ordered set.
It is easy to write all 14 relational operators by combinations of the symbols for not, less, greater and equal, so parsing this in a programming language would be easy.
This lack of awareness about partial order relations and the lack of support in most programming languages is very bad, because practical applications need very frequently partial orders instead of total orders.
For the floating-point numbers, the IEEE standard specifies 2 choices. You can either use them as a totally-ordered set, or as a partially-ordered set.
When you encounter NaNs as a programmer, that is because you have made the choice to have partially-ordered FP numbers, so you are not allowed to complain that this is an odd behavior, when you have chosen it. Most programmers do not make this choice consciously, because they just use the default configuration of the standard library, but it is still their fault if the default does not do what they like, but nonetheless they have not changed the default settings.
If you do not want NaNs, you must not mask the invalid operation exception. This is actually what the IEEE standard recommends as the default behavior, but lazy programmers do not want to handle exceptions, so most libraries choose to mask all exceptions in their default configurations.
When invalid operations generate exceptions, there are no NaNs and the FP numbers are totally ordered, so the 6 relational operators behave as naive programmers expect them to behave.
If you do not want to handle the invalid operation exception and you mask it, there is no other option for the CPU than to use a special value that reports an invalid operation, and which is indeed not-a-number. With not-numbers added to the set of FP numbers, the set becomes a partially-ordered set and all relational operators must be interpreted accordingly.
If you use something like C/C++, with only 6 relational operators, then you must do before any comparison tests to detect any NaN operand, because otherwise the relational operators do not do what you expect them to do.
In a language with 14 relational operators, you do not need to check for NaNs, but you must choose carefully the relational operator, because for a partially-ordered set, for example not-less is not the same with greater-or-equal (because not-less is the same with greater-or-equal-or-unordered).
If you do not expect to do invalid operations frequently, it may be simpler to unmask the exception, so that you will never have to do any test for NaN detection.
Fun fact - in C++ std::sort has undefined behavior, and can crash[1], if you try to sort a container with NaNs in it.
[1] https://stackoverflow.com/questions/18291620/why-will-stdsor...
IEEE 754 prescripes, for better or worse, that any mathematical comparison operator (==, <, > ….) involving at least one NaN must always return false, including comparison against itself. This is annoying for something like dictionaries or hashtables. C# has a solution: if you call a.Equals(b) on two floats a and b, it will return true also if both are NaN. I think this is a cool solution: it keeps the meaning of math operators the same identical with other languages, but you still have sensible behavior for containers. I believe this behavior is copied from Java.
I guess because the hash of an instance stays consistent (which is used to retrieve the value from the dict). The `__eq__` method must disregard the hash and return False for all nans.
But the hash alone shouldn't be enough to match the key. Isn't an equality check also needed to avoid a false positive? That's the idea behind a hash table, as I understand it. (I'm not a Python programmer.)
That equality check also considers object identity first:
I'm pretty sure that this is meant as an optimization.(But it does have to find the instance via the hash lookup first. This won't work if you e.g. return a random number from `__hash__`.)
> Last week in the Python Discord we had an unusual discussion about a Python oddity.
Oh, I missed it. But yes, this is more to do with NaN than Python.
> But, of course, you can't actually get to those values by their keys: ... That is, unless you stored the specific instance of nan as a variable:
Worth noting that sets work the same way here, although this was glossed over: you can store multiple NaNs in a set, but not the same NaN instance multiple times. Even though it isn't equal to itself, the insertion process (for both dicts and sets) will consider object identity before object equality:
And, yes, the same is of course true of `collections.Counter`.If I could change one thing in computing, it'd be how SQL handles NULL. But if I got a second thing, it'd be how IEEE handles NaN. I probably wouldn't even allow NaN as a representation. If some mathematical operation results in what would be NaN, I'd rather force the programming language to throw some sort of interrupt or exception. Much like what happens when you divide an integer by 0. Heck, I'd probably even stop infinity from being represented with floats. If someone did 1/0 or 0/0, I'd interrupt rather than generating an INF or NaN.
In my experience, INF and NaN are almost always an indicator of programming error.
If someone want's to programmatically represent those concepts, they could do it on top of and to the side of the floating point specification, not inside it.
Floating-point infinity is actually really useful in Python because you can easily and efficiently compare it to Python's arbitrary-size integers.
For what purpose? When does this come up?
I can also compare infinity to Java's BigDecimal values but I fail to see what I'd want to or need to.
Reminds me of this classic 4chan thread which started with an absurd-sounding comparison operator and ended with NaN semantic revelations.
https://archive.tinychan.net/read/prog/1176222557
This is also why Rust has separate PartialEq and Eq traits - the latter is only available for types that don't have weird not-self-equal values like floating point NaNs or SQL NULLs. If you lie to Rust and create a wrapper type over f32 or f64 that has Eq, then you'd get unindexable NaN keys that just sit in your hashmap forever.
The real surprise to me is that Python can index NaN keys sometimes, at least by reference to the original NaN. I knew CPython does some Weird Shit with primitive values, so I assume it's because the hashmap is comparing by reference first and then by value.
It would be more satisfying to learn why hash of nan is not guaranteed to be the same. It feels like a bug.
For binary operations, NaN values compare as unordered.
The IEEE 754 Specification requires that >,<,= evaluate to False.
Saying that two incomparable objects become comparable let alone gain equally would break things.
We use specific exponents and significands to represent NaNs but they have no numerical meaning.
I am actually surprised python got this correct, often NaN behavior is incorrect out of convenience and causes lots of issues and side effects.
At the standards level, NaN payload propagation isn't guaranteed, regardless of any other issues.
> payload propagation isn't guaranteed
Yes and no:
`If an operation has a single NaN input and propagates it to the output, the result NaN's payload should be that of the input NaN (this is not always possible for binary formats when the signaling/quiet state is encoded). If there are multiple NaN inputs, the result NaN's payload should be from one of the input NaNs; the standard does not specify which.'
The hash is the same. But a hash set has to use == in case of equal hashes (to avoid collisions).
It's not always the same:
maybe it's that multiple bit patterns can be NaN and these are two different ones? In IEEE-754, a number with all the exponent bits set to 1 is +/-infinity if the fraction bits are all zero, otherwise it's NaN. So these could be values where the fractions differ. Can you see what the actual bits it's setting are?
Yes. The CPython hash algorithm for floats (https://github.com/python/cpython/blob/main/Python/pyhash.c#...) special-cases the non-finite values: floating-point infinities hash to special values modeled on the digits of pi (seriously! See https://github.com/python/cpython/blob/main/Include/cpython/...), and NaNs fall through ultimately to https://github.com/python/cpython/blob/main/Include/internal... which is based on object identity (the pointer to the object is used rather than its data).
Probably just due to encoding. NaN is all 1s for the exponent and non-zero mantissa, so that's 2^23 - 1 possible values for f32
NaN == NaN is truly a perversion of equality.
It makes little sense that 1/0 is SIGFPE, but log(-5) is NaN in C.
And the same is true for higher level languages, and their error facilities.
What a mess.
Makes perfect sense .
NaN is a special type indicating one can't reason about it normal way.
It is an unknown or value that can't be represented.
When comparing, think of it like comparing two bags of unknown amount of apples.
One bag has NaN count of apples
Other bag has NaN count of apples
Do the two bags have equal number of apples?
I wish all languages used nulls the way SQL does.
Respectfully, I disagree.
If NaNs were meant to represent unknown quantities, then they would return false for all comparisons. But NaN != NaN is true. Assuming that two unknowns are always different is just as incorrect as assuming that they're always the same.
I'd also push back on the idea that this behavior makes sense. In my experience it's a consistent source of confusion for anyone learning to program. It's one of the clearest violations of the principle of least astonishment in programming language design.
As others have noted, it makes conscientious languages like Rust do all sorts of gymnastics to accommodate. It's a weird edge case, and imo a design mistake. "Special cases aren't special enough to break the rules."
Also, I think high level languages should avoid exposing programmers to NaN whenever possible. Python gets this right: 0/0 should be an error, not a NaN.
As a standard for floating point representation and computation, IEEE 754 solved multiple long-standing serious problems better than anything that came before it. I don't think its sensible to judge it with a PL design lens like "principle of least astonishment"; certainly not as if IEEE754 is a peer to Rust or Python. Or, you could learn about the surprise, frustration, and expense involved in trying to write portable numeric code in the 1970s, prior to IEEE 754:
https://people.eecs.berkeley.edu/~wkahan/ieee754status/754st...
I like this justifiation of NaN != NaN; it emphasizes that NaN has representional intent, more than just some bit pattern.
We take for granted that (except for things like x86 extended precision registers) floating point basically works the same everywhere, which was the huge victory of IEEE 754. It easy to lose sight of that huge win, and to be ungrateful, when one's first introduction to IEEE 754 are details like NaN!=NaN.
Python is extra annoying though with refusing to support division through zero the way other programming languages with IEEE floats do (i.e. output inf or nan instead of throwing an exception), even though it has no problem doing things like float('inf') / float('inf') --> nan. It specifically does it for division through zero as if it wants to be a junior grade calculator just for this one thing. They could at least have fixed this when breaking backwards incompatibility from python2 to 3...
Nah. Python gets it right; all high level languages should operate this way. Division by zero is a bug 90% of the time. Errors should never pass silently. Special cases aren't special enough to break the rules.
IEEE floats should be a base on which more reasonable math semantics are built. Saying that Python should return NaN or inf instead of throwing an error is like saying that Python should return a random value from memory or segfault when reading an out-of-bounds list index.
In most languages, `x: float = 0` involves an implicit conversion from int to float. In Python, type annotations have no impact on runtime behavior, so even though the type checker accepts this code, `type(x)` will be `int` -- python acts as if `int` was a subtype of `float`.
It would be weird if the behavior of `1 / x` was different depending on whether `0` or `0.0` was passed to a `x: float` parameter -- if `int` is a subtype of `float`, then any operation allowed on `float` (e.g. division) should have the same behavior on both types.
This means Python had to choose at least one:
1. division violates the liskov substitution principle
2. division by zero involving only integer inputs returns NaN
3. division by zero involving only float inputs throws exception
4. It's a type error to pass an int where a float is expected.
They went with option 3, and I think I agree that this is the least harmful/surprising choice. Proper statically typed languages don't have to make this unfortunate tradeoff.
C does different things for 0.0 / 0.0 and 0 / 0 and it's not that weird to deal with (well it has other issues like it being platform dependent what happens with this). JS has no problem with it either (0.0 / 0.0 gives nan, 0n / 0n gives exception since it are integers).
Python is the only language doing this (of the ones I use at least).
I don't think the notation `x: float = 0` existed when it was new by the way so that can't be the design reason?
since python seems to handle integer through integer divisions as float (e.g. 5 / 2 outputs 2.5), 0 / 0 giving nan would seem to be expected there
> liskov substitution principle
that would imply one is a subtype of another, is that really the case here? there are floats that can't be represented as an integer (e.g. 0.5) and integers that can't be represented as a double precision float (e.g. 18446744073709551615)
Python chose, quite some time ago, not to follow C's lead on division: PEP 238 – Changing the Division Operator (2001) [1]
The rationale is basically that newcomers to Python should see the results that they would expect from grade school mathematics, not the results that an experienced programmer would expect from knowing C. While the PEP above doesn't touch on division by zero, it does point toward the objective being a cohesive, layman-friendly numeric system.
[1]: https://peps.python.org/pep-0238/
Just use numpy if you want to do math. Seriously.
There's no non-confusing option for comparisons. You have two invalid values, but they aren't necessarily the same invalid value. There are multiple operations that can produce NaN.
It's a sentinel value for an error. Once you have an error, doing math with the error code isn't sensible.
There are no non-confusing options, but some of those are still clearly worse than others.
What should sorted([3, nan, 2, 4, 1]) give you in Python?
A) [1, 2, 3, 4, nan] is an good option
B) [nan, 1, 2, 3, 4] is an good option
C) An error is an good option
D) [3, nan, 1, 2, 4] is a silly, bad option. It's definitely not what you want, and it's quiet enough to slip by unnoticed. This is what you get when Nan != NaN
NaN == NaN is wrong. NaN != NaN is wrong, unintuitive, and breaks the rest of your code. If you want to signal that an operation is invalid, then throw an error. The silently nonsensical semantics of NaN are the worst possible response
...reminds me that object + object is NaN:
see https://www.destroyallsoftware.com/talks/watPS: Wait for it ... Watman! =8-)
NaN is weird? No, NaN is normal*, NaN PAYLOADS are weird: https://anniecherkaev.com/the-secret-life-of-nan
*This is false, NaN is weird, though maybe it needs to be. It is nowhere written arithmetic on computers must be straightforward.
Actually none of the floating point value are normal :D
Although Almost All† Real Numbers are Normal, none of the floats are normal, and nor are most numbers any regular person would think of (possible exceptions Pi and Euler's Number which are conjectured to be Normal although it is unproven)
† Almost All is a term of art in mathematics, remember there are uncountably many real numbers, so the fact the Normals are also uncountable puts them at least in the right ballpark, unlike the rationals.