structural_observer.eigs

# Structural drill — observer predicates
# Minimal patterns for converged, stable, improving, report, observe.
# Every line parses. Every loop has a body.

# ---- report ----
x is 10.0
status is report of x
print of status

y is 5.0
y is 4.0
y is 3.0
print of (report of y)

# ---- observe ----
z is 1.0
z is 0.5
z is 0.25
state is observe of z
print of state

# ---- loop while not converged (inside function for fresh scope) ----
define shrink as:
    sh_v is n
    loop while not converged:
        sh_v is sh_v * 0.9
    return sh_v

result is shrink of 100.0
print of result

# ---- Newton square root ----
define sqrt_newton as:
    sn_x is n * 1.0
    sn_g is sn_x / 2.0
    if sn_g < 1.0:
        sn_g is 1.0
    loop while not converged:
        sn_g is (sn_g + sn_x / sn_g) / 2.0
    return sn_g

r is sqrt_newton of 16.0
print of r

# ---- Fixed point ----
define fixed_point as:
    fp_x is n * 1.0
    loop while not converged:
        fp_x is (fp_x + 1.0 / fp_x) / 2.0
    return fp_x

fp is fixed_point of 2.0
print of fp

# ---- Status-driven loop at top level ----
value is 50.0
status is "start"
loop while status != "converged":
    value is value * 0.85
    status is report of value

print of value

# ---- Detect stable band ----
signal is 80.0
status is "start"
found_stable is 0
loop while status != "converged":
    signal is signal * 0.9
    status is report of signal
    if status == "stable":
        if found_stable == 0:
            found_stable is 1

print of signal

# ---- Gradient descent with convergence ----
define descend as:
    dg_x is n * 1.0
    dg_target is 3.0
    loop while not (equilibrium of dg_x):
        dg_grad is 2.0 * (dg_x - dg_target)
        dg_x is dg_x - 0.1 * dg_grad
    return dg_x

d is descend of 10.0
print of d