ArbReals

---

other, sometimes overlapping, software development is with ArbFloats

---

current flows are in the wiki and on gitter

######real intervals with floating point bounds (uses the Arb library)

######                                                                                                                          To participate, email a moderator.

visit the Wiki to see current todos/thoughts/perspecitves
    and to use the whiteboards -- participate to edit them

on gitter

---

##Use (interval valued)

using ArbReals

five = midpoint_radius(5.0, 0.0)
# 5.00000000000000

tanh(five)
# [0.999909204262595 +/- 1.32e-16]

showall(ans)
# [0.9999092042625951312109904475344730 +/- 2.59e-35]

# ----- 

exp1interval = midpoint_radius( exp(1), eps(exp(1)) )
# [2.71828182845905 +/- 5.36e-15]

showall(exp1interval)
# [2.718281828459045090795598298427649 +/- 4.45e-16]

showcompact(exp1interval)
# [2.7182818 +/- 2.85e-8]

midpoint(exp1interval)
# 2.71828182845905

radius(exp1interval)
# 4.44089210677243e-16

a = agm(exp1interval, inv(exp1interval))
# [1.25683135663827 +/- 4.99e-15]

showall(ans)
# [1.256831356638274777545517650643936 +/- 2.06e-16]

showall(lowerbound(a))

showall(upperbound(a))
# 1.256831356638274982902060394630015

showall(lowerbound(ans))
# 1.256831356638274572188974906657858

ToDo

1. find a fast way to express (midpoint & radius) as a most informed value
   >> See ArbFloats' string functions for a way that works, faster would help.

2. make library's fetch & build independant of Nemo

3. wrap and export standard interval predicates and operators

4. define missing math functions (csc, sec, csch, sech, acsc, asec..)