来自whatis.techtarget.com:

In information technology, a boilerplate is a unit of writing that can be reused over and over without change. By extension, the idea is sometimes applied to reusable programming as in "boilerplate code." The term derives from steel manufacturing, where boilerplate is steel rolled into large plates for use in steam boilers. The implication is either that boilerplate writing has been time-tested and strong as "steel," or possibly that it has been rolled out into something strong enough for repeated reuse.

编程之外:

样板可以与特定类型的模板进行比较 被认为是一个填空样板。一些典型的 样板文件通常包括:任务陈述、安全警告 使用的安装程序、版权声明和责任 免责声明。

以我作为程序员的经验来看，合适的样板代码通常是一堆你开始的代码，这些代码不太大，也不太复杂，不能称为框架。

HTML5 Boilerplate就是一个典型的例子。

Joshua Bloch有一个关于API设计的演讲，涵盖了糟糕的API如何使样板代码成为必要。(参考样板文件的第46分钟，今天听这个)

它是可以被许多应用程序/上下文使用的代码，只需很少或没有更改。

样板纸(Boilerplate)起源于20世纪初的钢铁工业。

“样板代码”是任何看似重复的代码，它们一次又一次地出现，以获得一些看起来应该简单得多的结果。

这是一个主观的定义。

这个术语来源于报纸行业的“标准版”:wiki

样板代码是指可以反复使用的一段代码。另一方面，任何人都可以说它是一段可重用的代码。

这个词实际上来自钢铁行业。

对于一点历史，根据维基百科:

在19世纪90年代，样板版实际上是铸造或冲压成金属，准备印刷，并分发给美国各地的报纸。直到20世纪50年代，成千上万的报纸从美国最大的供应商西部报业联盟那里收到并使用了这种样板文件。有些公司还以样板文件的形式发布新闻稿，这样它们就必须按书面形式打印出来。

根据维基百科:

In object-oriented programs, classes are often provided with methods for getting and setting instance variables. The definitions of these methods can frequently be regarded as boilerplate. Although the code will vary from one class to another, it is sufficiently stereotypical in structure that it would be better generated automatically than written by hand. For example, in the following Java class representing a pet, almost all the code is boilerplate except for the declarations of Pet, name and owner: public class Pet { private PetName name; private Person owner; public Pet(PetName name, Person owner) { this.name = name; this.owner = owner; } public PetName getName() { return name; } public void setName(PetName name) { this.name = name; } public Person getOwner() { return owner; } public void setOwner(Person owner) { this.owner = owner; } }

In practical terms, boilerplate code is the stuff you cut-n-paste all over the place. Often it'll be things like a module header, plus some standard/required declarations (every module must declare a logger, every module must declare variables for its name and revision, etc.) On my current project, we're writing message handlers and they all have the same structure (read a message, validate it, process it) and to eliminate dependencies among the handlers we didn't want to have them all inherit from a base class, so we came up with a boilerplate skeleton. It declared all the routine variables, the standard methods, exception handling framework — all a developer had to do was add the code specific to the message being handled. It would have been quick & easy to use, but then we found out we were getting our message definitions in a spreadsheet (which used a boilerplate format), so we wound up just writing a code generator to emit 90% of the code (including the unit tests).