# Replicating pi(x) from Era Ordering with a Script
The numerical question is simple: if integers are bounded to the earliest era
in which their generators are available, what does the cumulative admitted set
look like when it is sorted back into the usual order on `N`?
The script for this experiment now lives at
[`C:\Users\an\src\siran\writing\.scripts\tools\era_order.py`](C:\Users\an\src\siran\writing\.scripts\tools\era_order.py).
It now works era by era rather than through a fixed numerical cutoff. On each
run it:
1. loads the cached era state, if any;
2. saves a rolling `latest` image of the current era-truncated counting plots;
3. reports the known eras and the size of the admitted set;
4. asks how many more eras to compute, together with a timing estimate based on
the most recent era;
5. computes those eras, updates the cache, and saves both per-era snapshots and
a refreshed `latest` plot;
6. can optionally save checkpoint plots during materialized eras.
When the admitted right endpoint becomes too large to read comfortably, the
script automatically switches the global horizontal axis to `log10(N)` so late
eras remain visible and very large integer endpoints do not break plotting.
The plotted function is not the classical prime-counting function `pi(N)`.
It is the era-truncated version
$$
\pi_E(N)=\#\{p\le N : p \text{ is prime and has been admitted through era } E\}.
$$
That distinction matters. The point of the experiment is to watch `\pi_E(N)`
approach the familiar jagged prime frontier as more eras are added, not to
pretend that a low-era truncation is already the full classical `\pi(N)`.
Each saved image now contains two complementary panels:
1. a dense local staircase, with one plotted point for each successive
integer in a finite window;
2. the global compressed frontier of `\pi_E(N)` across the whole admitted
range.
By default the dense local window runs from `1` through the current era, so the
upper panel shows the recovered staircase in the region where it is already
classical. A larger dense window can be requested explicitly with
`--dense-xmax`, in which case the plot plateaus after the currently admitted
primes.
It uses the current era rule:
$$
\tau(1)=1,
$$
and for $n>1$,
$$
\tau(n)
=
\min\left(
\kappa(n),
\min_{\substack{ab=n\\1
---
- [Preferred Frame Writing on GitHub.com](https://github.com/siran/writing)
(built: 2026-03-25 18:30 EDT UTC-4)