Just a question: Anyone know why DoubleRange does ...
# stdlibv
vmichalak
08/07/2018, 1:04 PMJust a question: Anyone know why DoubleRange does not exists in the StdLib ?
c
Czar
08/07/2018, 1:06 PMbecause Double is neither discrete nor exact type.
Czar
08/07/2018, 1:07 PME.g. what are the steps between 1.0 and 2.0? between 1.0 and 1.1? 1.0 and 1.0000001?
v
vmichalak
08/07/2018, 1:08 PMYou can specify the step for Int and Long, why not for Double
c
Czar
08/07/2018, 1:08 PMbecause step involves math operations on a double, which are not always deterministic exact.
k
karelpeeters
08/07/2018, 1:09 PM*exact
✔️ 1
karelpeeters
08/07/2018, 1:09 PMIEEE floating point math is definitely deterministic.
s
spand
08/07/2018, 1:11 PMSeems like an odd reason. Has there been an official response to this ?
c
Czar
08/07/2018, 1:12 PMhow is that an odd reason?
s
spand
08/07/2018, 1:13 PMBecause if one has chosen to work with Double in using a DoubleRange why should its inexactness not carry over?
spand
08/07/2018, 1:15 PMIf anything I would expect such a type to handle the common gotchas that probably would arise from this instead of redoing it poorly as was done in slack maybe two months ago
c
Czar
08/07/2018, 1:16 PMI'd say because it's error-prone and bad style in most situations. If one is in a situation where that would be a good idea, nothing prevents them from implementing DoubleRange themselves.
Czar
08/07/2018, 1:21 PMActually stdlib has
ClosedDoubleRangeClosedFloatingPointRange👍 1
s
spand
08/07/2018, 1:41 PM