关于「自校验编译器:码位与整条流水线」,《计算著作学与自校验专著》给出的定量结论是什么?
结论:第六章清点印出来像汉字的码位(全部 214 个康熙部首),并计算整条流水线把含 81 个错误数字的草稿降到预期 1.322 个,全无错误的机会是 0.2665,而不是零。
数据口径:书中的读者、书、错误率与成本均为示意模型;全部计算由书中程序生成;本书不是法律、财务或出版建议。 书中以标注框区分已发表的证据、可运行的 Python 计算(16 段程序)、原典与作者的假说。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.2665, 1.322
| stages | errors left | P(zero errors) |
|---|---|---|
| draft | 81.000 | 0.0000 |
| A | 12.109 | 0.0000 |
| G | 20.388 | 0.0000 |
| GB | 14.313 | 0.0000 |
| AG | 3.048 | 0.0475 |
| AGB | 2.140 | 0.1177 |
| GBT | 8.845 | 0.0001 |
| AGBT | 1.322 | 0.2665 |
| raw rate 0.18 -> per-number rate after AGBT 0.00294 |
查看完整验算程序
# Chapter 6. p12: From an 18% draft to how many errors? A 450-number
# draft at 0.18 holds 81 wrong numbers, mixed 0.60 invented, 0.25 slips,
# 0.15 swaps. Stages, each applied to what the last one left:
# A three human passes, detectability Beta(0.7, 0.3) (miss 0.1495)
# G chapter-local whitelist (escape: invented 0.122, slip 0.114,
# swap 1.000; values from p10)
# B bilingual set equality: a swap made in one language only is caught
# if it changes the set of numbers; assume half of swaps are
# one-language and those are caught
# T numbers typed from the program output by a template: swaps of
# that kind cannot occur; assume 0.9 of swaps are removed
# Expected errors left and P(zero errors) under a Poisson count.
import math
from math import exp
DRAFT = 81.0
MIX = {"invented": 0.60, "slip": 0.25, "swap": 0.15}
GATE = {"invented": 0.122, "slip": 0.114, "swap": 1.0}
def run(stages):
e = {k: DRAFT * v for k, v in MIX.items()}
for s in stages:
if s == "A":
e = {k: v * 0.1495 for k, v in e.items()}
elif s == "G":
e = {k: v * GATE[k] for k, v in e.items()}
elif s == "B":
e["swap"] *= 0.5
elif s == "T":
e["swap"] *= 0.1
return math.fsum(e.values())
if __name__ == "__main__":
print(" stages errors left P(zero errors)")
for st in ("", "A", "G", "GB", "AG", "AGB", "GBT", "AGBT"):
left = run(st)
print(" {:<14} {:11.3f} {:14.4f}".format(
st or "draft", left, exp(-left)))
print(" raw rate 0.18 -> per-number rate after AGBT {:.5f}".format(
run("AGBT") / 450))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《计算著作学、先证伪后立界修辞与自校验专著编译器》第 6 章「自校验编译器:码位与整条流水线」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-101 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026101,
author = {Fan, Yafu},
title = {Computational Bibliometrics, Falsification-First Rhetoric and Self-Verifying Monograph Compilers},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}