Oct. 11th, 2008
Читаю "Сложность задач и эффективность методов оптимизации" Немировского и Юдина в английском переводе и натыкаюсь на чудесное:
The classification of ... [optimization] problems according to their 'smoothness' is justified not by the fact that in practice we often encounter convex problems which are not Lipschitz, or Lipschitz-convex problems which are differentiable only a few times. Such a justification would be, so to say, the 'pure water of demagogy' -- rank deception by deliberate distortion of the facts. For the mere fact of a high degree of smoothness does not, as a rule, arouse any doubts.Вот как надо писать научные книги!