Examples

The repository includes executable examples under examples/:

• rump.jl — Rump polynomial instability benchmark
• catastrophic_cancellation.jl — cancellation amplification
• muller_sequence.jl — recurrence sensitivity
• logistic_map.jl — chaotic iteration with significance decay
• branching_instability.jl — stochastic branch divergence
• summation_kahan.jl — naive vs Kahan summation under DSA
• hilbert_solve.jl — linear-system conditioning stress test

Run any example:

julia --project=. examples/rump.jl

Typical usage pattern

using DSArith

val = (ds(1.0) + ds(1e-16)) - ds(1.0)
println(stochastic_string(val))
println(report(val))

Typical workflow:

1. replace key Float64 values with ds(...),
2. compute as usual,
3. inspect report(...) and diagnostics,
4. use low significance / @.0 as a trigger for algorithmic improvements.