< and > breaks interval?

[Moelf]

julia> f(x) = x>0.5
f (generic function with 1 method)

julia> f(0..1)
false

julia> f(1..2)
true

is this somehow expected according to IEEE?

[OlivierHnt]

I believe so; the reference for this is Table 10.3 of the IEEE Standard 1788-2015.

As a side note: inequalities look broken on the 1.0-dev branch

julia> 0..1 > 0.5
true

julia> 0.5 < 0..1
false

[Kolaru]

Currently < implements the strictLess function from the standard, as @OlivierHnt pointed out.

However, the symbol is only a suggestion, so we have the freedom to do whatever we want with it. For interoperability with the julia ecosystem, I think the construct should just error. Otherwise, some julia code silently fails (see #165). It is one of the reason that lead to the creation of NumberIntervals.jl too.

It is one of the things that would be great to have sorted out for 1.0.

So yes, intervals break < and >.