我认为您需要检查一些您作为开发人员遇到过的问题，当您因为另一个设计更改而不得不第十次修改代码时，您会感到非常紧张。你可能有一个项目清单，你觉得有很多返工和痛苦。

From that list you can derive the scenarios that the Design Patterns intend to solve. Has there been a time where you needed to perform the same series of actions on different sets of data? Will you need to be able to future capability to an application but want to avoid reworking all your logic for existing classes? Start with those scenarios and return to the catalog of patterns and their respective problems they are supposed to solve. You are likely to see some matches between the GoF and your library of projects.

我不知道哪本书最好，但纯粹主义者可能会说《设计模式:可重用面向对象软件的元素》

至于我个人最喜欢的，我喜欢O'Reilly出版的Head First Design Patterns。它是用对话的口吻写的，很吸引我。当我阅读它时，我同时检查了我的源代码，看看它是否适用于我所阅读的内容。如果有，我就重构。我就是这样学会责任链的。

练习-练习-练习。

我领导了一些设计模式讨论小组(我们的网站)，并阅读了5或6本模式书籍。我建议从《头部优先设计模式》一书开始，参加或发起一个讨论小组。《Head First》一开始可能看起来有点像Hasboro，但大多数人在读了一两章后就喜欢上了它。

Use the outstanding resource - Joshua Kereivisky's A Learning Guide to Design Patterns for the pattern ordering and to help your discussion group. Out of experience the one change I suggest to the ordering is to put Strategy first. Most of today's developers have experienced some good or bad incarnation of a Factory, so starting with Factory can lead to a lot of conversation and confusion about the pattern.This tends to take focus off how to study and learn patterns which is pretty essential at that first meeting.

你试过《四人帮》这本书吗?

设计模式:可重用面向对象软件的元素

我发现，要理解或理解某些模式的好处有点困难，除非理解它们解决的问题和其他(更糟糕的)实现问题的方式。

除了GOF和POSA的书，我还没有真正读过，所以我不能给你其他的建议。实际上，您只需要了解问题领域，我认为许多缺乏经验的开发人员可能无法欣赏模式的好处。这不是对他们的轻视。当一个人不得不首先与糟糕的选择作斗争时，接受、理解和欣赏好的解决方案要容易得多。

祝你好运

对于这样一个老问题，我只有两分钱

有些人已经提到，实践和重构。我认为学习模式的正确顺序是:

学习测试驱动开发(TDD) 学习重构 学习模式

大多数人忽略了1，许多人相信他们可以做2，几乎每个人都直接去做3。

对我来说，提高软件技能的关键是学习TDD。这可能是一个痛苦而缓慢的编码过程，但首先编写测试肯定会让您对代码进行很多思考。如果一个类需要太多样板文件或者很容易崩溃，你很快就会注意到不好的味道

TDD的主要好处是你不再害怕重构你的代码，并迫使你编写高度独立和内聚的类。如果没有一套好的测试，触摸没有损坏的东西就太痛苦了。有了安全网，你将真正冒险对你的代码进行巨大的更改。这是你真正开始从实践中学习的时刻。

Now comes the point where you must read books about patterns, and to my opinion, it is a complete waste of time trying too hard. I only understood patterns really well after noticing I did something similar, or I could apply that to existing code. Without the safety tests, or habits of refactoring, I would have waited until a new project. The problem of using patterns in a fresh project is that you do not see how they impact or change a working code. I only understood a software pattern once I refactored my code into one of them, never when I introduced one fresh in my code.