Rewrite the NLOS post to reflect what microwaveprop has learned:
M-refractivity and dM/dh as the actual duct discriminant, the four
duct taxonomy with radiation ducts as the North Texas mechanism,
binary duct detection as a coin flip, low pressure beating high,
shallow boundary layer winning, PWAT sweet spot, seasonal pattern
(spring worst, summer best), troposcatter as the floor, and real
numbers for the 65 km QTH-to-W5HN/B path.
New interactive scenes: N vs M side-by-side refractivity profile,
pressure/HPBL/binary-duct bar chart, band-vs-dawn enhancement,
PWAT sweet spot with draggable cursor, monthly ducting probability,
and a 10 GHz link-budget decomposition showing how the 36 dB
diffraction gap gets closed.
Polish pass across all scenes: softer palette, rounded stroke
caps/joins, prettier slider styling in scene.css and nlos.css,
signal-pulse animations on NF cascade / single-stage / LNA
placement, trail persistence on the ray-trace scene, animated
glints along trapped rays in the duct scene, and animated TCP
sequence diagrams.
Add local-serve preview mode so future-dated posts show up on
zola serve but stay hidden on the deployed build.
Filter future-dated posts out of blog/homepage listings and add a
daily 05:05 UTC rebuild so posts go live on their date automatically.
New interactive post on BJT amplification with five canvas scenes
(valve, load line, transfer curve, amplifier in action, gain).
Shared scene.css for future interactive posts. Post is dated
2026-04-23 and will appear in listings once the date arrives.
Also drop IDEAS.md with candidate topics for future posts.
New post "why-signals-bend" with ten hand-rolled canvas scenes
(LOS, Earth curvature, Fresnel, knife-edge and rounded diffraction,
refractivity profile, ray tracing, duct trapping, duct types,
and the real 65 km QTH-to-W5HN/B path with 4/3-Earth bulge).
Also drop QSO/em-dash conventions from prop1 and the
"written by humans" footer line.
