bt。

我认为自20世纪80年代以来发明的最好的想法将是我们不知道的。要么是因为它们很小，无处不在，以至于不引人注意，要么是因为它们的受欢迎程度还没有真正起飞。

前者的一个例子是单击并拖动以选择文本的一部分。我相信这是1984年首次出现在麦金塔电脑上。在此之前，您有单独的按钮用于选择选择的开始和结束。相当繁重。

后者的一个例子是(可能是)可视化编程语言。我不是说像hypercard，我是说像Max/MSP, Prograph, Quartz Composer, yahoo pipes等。目前它们确实是小众的，但我认为，除了思想分享之外，没有什么能阻止它们像标准编程语言一样具有表现力和强大的功能。

可视化编程语言有效地加强了引用透明性的函数式编程范式。这对于代码来说是一个非常有用的属性。他们执行这一点的方式也不是人为的——这只是由于他们使用的比喻。

VPL让那些本来不会编程的人也能编程，比如有语言障碍的人，比如阅读困难的人，甚至只是需要简单节省时间的门外汉。专业程序员可能会对此嗤之以鼻，但就我个人而言，我认为如果编程成为一种真正无处不在的技能，就像识字一样，那就太好了。

就目前来看，VPL只是一个小众的兴趣，还没有真正成为主流。

我们应该做些什么不同的事情

all computer science majors should be required to double major- coupling the CS major with one of the humanities. Painting, literature, design, psychology, history, english, whatever. A lot of the problem is that the industry is populated with people that have a really narrow and unimaginative understanding of the world, and therefore can't begin to imagine a computer working any significantly differently than it already does. (if it helps, you can imagine that I'm talking about someone other than you, the person reading this.) Mathematics is great, but in the end it's just a tool for achieving. we need experts who understand the nature of creativity, who also understand technology.

But even if we have them, there needs to be an environment where there's a possibility that doing something new would be worth the risk. It's 100 times more likely that anything truly new gets rejected out of hand, rather viciously. (the newton is an example of this). so we need a much higher tolerance for failure. We should not be afraid to try an idea which has failed in the past. We should not fully reject our own failures- and we should learn to recognize when we have failed. We should not see failure as a bad thing, and so we shouldn't lie to ourselves or to others about it. We should just get used to it, because it is just about the only constant in this ever changing industry. Post mortems are useful in this regard.

One of the more interesting things, about smalltalk, I think, was not the language itself, but the process that was used to arrive at the design of smalltalk. The iterative design process, going through many many revisions- But also very carefully and critically identifying the flaws of the existing system, and finding solutions in the next one. The more perspectives, and the broader the perspectives we have on the situation, the better we can judge where the mistakes and problems are. So don't just study computer science. Study as many other academic subjects as you can get yourself to be interested in.

在主流计算中，有一件事没有改变，那就是分级文件系统。在我看来，这是一种耻辱，因为在20世纪80年代末和90年代，一些工作已经完成，以设计更适合现代面向对象操作系统的新型文件系统——那些从头开始就是面向对象的操作系统。

OO操作系统倾向于具有可扩展且灵活的平面对象存储。我认为EROS项目就是围绕这个想法建立的;PenPoint操作系统是20世纪90年代的面向对象操作系统;Amazon S3当然是当代的平面对象存储。

在面向对象的平面文件系统中，至少有两个想法是我特别喜欢的:

The entire disk was essentially swap space. Objects exist in memory, get paged out when they are not needed, and brought back in when they are. There's no need for a hierarchical filesystem that's separate from virtual memory. Programs are "always running," in a sense. A flat file/object store allows content to be indexed and searched, rather than forcing the user to decide -- ahead of time -- where the content will live in relation to other content and what its name shall be. A hierarchical system could be built on top of the flat storage, but it's not required.

正如Alan Cooper在他的书《About Face》中所述，分层文件系统是一个拼凑物，是为20世纪60年代和70年代内存和磁盘存储有限的计算机设计的。可悲的是，Windows和Unix的流行保证了分层文件系统的主导地位。

采用面向对象。

这个想法早前就出现了(例如Simula)，但在20世纪90年代成为主流。(恕我直言，它最大的好处之一是在开发人员之间提供了一个通用的词汇表，所以它的广泛采用使它更有价值。)

Damas-Milner type inference (often called Hindley-Milner type inference) was published in 1983 and has been the basis of every sophisticated static type system since. It was a genuinely new idea in programming languages (admitted based on ideas published in the 1970s, but not made practical until after 1980). In terms of importance I put it up with Self and the techniques used to implement Self; in terms of influence it has no peer. (The rest of the OO world is still doing variations on Smalltalk or Simula.)

类型推断的变化仍在上演;我最喜欢的变体是Wadler和Blott的解决重载的类型类机制，后来发现它为类型级别的编程提供了非常强大的机制。这个故事的结局还在书写中。

有目的的游戏——像Luis von Ahn和他的团队正在开发的集体智慧工具在1980年之前可能是一个梦想，但当时没有一个广泛部署的网络，可以容纳数百万人，并且需要(例如reCAPTCHA)来实现它。