关于「两种存续规律:为什么多数书褪色,少数书一直被读」,《计算著作学与自校验专著》给出的定量结论是什么?
结论:第一章说明「三年沦为废纸」是参数而不是定律:风险率上升的潮流书第 3 年只剩第一年注意力的 0.0055,风险率下降的参考书剩 0.3330,但起点大于 11.08 倍的潮流书永远不会被追上;引用的总量是有限的。
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: 0.0055, 0.3330, 11.08
year trend left ref left trend hazard ref hazard
1 0.3679 0.6147 1.500 0.400
3 0.0055 0.3330 2.598 0.240
10 0.0000 0.1165 4.743 0.100
30 0.0000 0.0359 8.216 0.037
attention-years over 30 years: trend 0.903, ref 4.257
all-time totals: trend 0.903, ref 10.000; ref can never catch a
trend book that starts more than 11.08 times larger
trend book starts 1 times larger: ref overtakes in year 0.5
trend book starts 2 times larger: ref overtakes in year 3.4
trend book starts 3 times larger: ref overtakes in year 7.7
trend book starts 5 times larger: ref overtakes in year 38.2查看完整验算程序
# Chapter 1. p01: Two kinds of book, two survival laws. A "trend" book's
# yearly attention follows a Weibull survival exp(-(t/lam)^k) with k > 1
# (decline speeds up); a "reference" book follows a Lomax (Pareto II) law
# (1 + t/s)^-a, whose hazard falls with age. Illustrative parameters.
# We print the share of first-year attention left, the hazard, and the
# year in which the reference book's cumulative attention overtakes a
# trend book that starts R times larger.
import math
from math import exp, gamma
LAM, K = 1.0, 1.5 # trend book: Weibull scale (years), shape
S, A = 2.0, 1.2 # reference book: Lomax scale (years), shape
def weib(t):
return exp(-(t / LAM) ** K)
def lomax(t):
return (1 + t / S) ** -A
def haz_w(t):
return K / LAM * (t / LAM) ** (K - 1)
def haz_l(t):
return A / (S + t)
def cum(f, T, n=30000):
h = T / n
return h * math.fsum(f((i + 0.5) * h) for i in range(n))
def crossover(r, horizon=200.0):
# first year in which cum(lomax) >= r * cum(weib)
t, step = 0.0, 0.05
while t < horizon:
t += step
if cum(lomax, t, 2000) >= r * cum(weib, t, 2000):
return t
return None
if __name__ == "__main__":
print(" year trend left ref left trend hazard ref hazard")
for t in (1, 3, 10, 30):
print(" {:4d} {:9.4f} {:8.4f} {:12.3f} {:10.3f}".format(
t, weib(t), lomax(t), haz_w(t), haz_l(t)))
cw, cl = cum(weib, 30), cum(lomax, 30)
print(" attention-years over 30 years: trend {:.3f}, ref {:.3f}"
.format(cw, cl))
tot_w = LAM * gamma(1 + 1 / K)
tot_l = S / (A - 1)
print(" all-time totals: trend {:.3f}, ref {:.3f}; ref can never"
" catch a".format(tot_w, tot_l))
print(" trend book that starts more than {:.2f} times larger"
.format(tot_l / tot_w))
for r in (1, 2, 3, 5):
x = crossover(r)
print(" trend book starts {} times larger: ref overtakes "
"{}".format(r, "never (200 y)" if x is None
else "in year {:.1f}".format(x)))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 1 章「两种存续规律:为什么多数书褪色,少数书一直被读」(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}
}