Streams for the damned
There are many ways of thinking about how asynchronous programs fit together. Streams are one way of composing asynchronous programs.
Varibales in a program change over time. Imagine each value a given variable takes throughout the course of a program. We can place these values on a timeline, with an x representing each time the value changed:
If values of a variables are events on a timeline (x's), then a stream is the timeline itself.
Streams tie events together - they create a temporal relationship between events. We can turn events into actions by transforming the events as they travel through the stream (e.g., from a click into a POST request; from a POST request into an element on the HTML page).
If you think about it, the state of a software interface is a function of the events (users' mashing buttons, data coming in from server) that happen over time. By describing streams of these events, where we transform these events into outcomes ("side-effects"), we can write interfaces in a more declarative style, where everything our interface does has a clear, causal relationship to something that happened during the application lifecycle.
(Still confused? Ok. Streams are like a TV channel. In some sense, TV channels are just the images that come over them; however, the idea of a channel can be a useful tool for reasoning about the images).
First, cd into this directory and
npm install. (You'll need node and npm. You'll also need to
npm install -g browserify).
Go through each file in order, 1, 2, 3, 4. For each file,
browserify [file] -o bundle.js and open index.html in your browser.
For example 4,
cd 4\ spire and
browserify app/index.js -o bundle.js