TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Topology of decisionmaking – are you secretly managing by metric?

5 点作者 jsweojtj将近 11 年前

3 条评论

dj-wonk将近 11 年前
&gt; Elementary topology shows that a good decision-making process is equivalent to managing by metric.<p>I have not checked the proof, but I think what the author means by &quot;equivalent&quot; is that it comes up with the same result, without regards to computational complexity, much less human and organizational psychology.<p>I would not expect the computational and informational costs to be the same for both. Shifting the order of computation may not matter, functionally speaking, but can affect efficiency in many cases. For example, some decisions are easier to make if you can defer gathering detailed information until later in the process: for example, you can prune low-scoring alternatives cheaply at the beginning.<p>Humans also often need to dig into a problem in order to explore the parameter space. This is roughly analogous to &quot;knowing it when you see it.&quot;
dj-wonk将近 11 年前
&gt; Yet in the real world, many people oppose metric-based management.<p>It seems to me that few organizations and even fewer people use mathematically rigorous decision-making. Some people realize this, some do not. I think some people oppose metric-based management out of self-preservation -- it involves hard thinking and some degree of formalization of their criteria. Specifying your criteria can box you in and give you less flexibility.
dj-wonk将近 11 年前
Question: the article does not directly mention the &quot;independence of irrelevant alternatives&quot; (IoIA) concept: <a href="https://en.wikipedia.org/wiki/Independence_of_irrelevant_alternatives" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Independence_of_irrelevant_alt...</a><p>I wonder if the author&#x27;s assumptions imply IoIA. I am inclined to think so, but have not proved it.