Differences between Node environment and browser javascript environment

Question

I've always been a bit annoyed that there are two major realms of javascript projects -- Node and "the browser" -- and while most browser JS can be easily run inside Node with a couple libraries for DOM stuff if needed, porting Node stuff to the browser is usually an afterthought.

This all seems like a lot of wasted energy on the part of developer communities, which could be alleviated by all JS developers just developing for the "least common denominator" (the browser) and using various shims to use features only available in Node or other JS environments besides the plain old browser.

This would not only cut out a lot ecosystem cruft and make development-in-the-browser more realistic, it also makes it commonplace to give the browser superpowers…Look for example at browserver, which sets up an http server inside the browser, but because the browser cannot actually accept http requests, uses websockets to talk to a proxy Node server that can.

So I want to ask, what are the real technical constraints of a web browser's javascript environment versus Node? I thought Node was just "a javascript environment, plus http server and local filesystem, minus the DOM and the chrome". Are there technical reasons why developers could not potentially move to the approach I described above, developing for the browser JS environment (does this have an official name?) and using shims for Node?

Answer 1

Code that runs on the client usually have very different goals from the code that runs on the server. However when it makes sense to use some libarary's features in both environments there are a lot of them that are defined using a universal AMD form which makes them platform independent (e.g. Q).

The major difference between both environments is that one is subject to rigorous security policies and restrictions (browser) while the other isin't. The browser is also an untrustable environment for security-related operations such as enforcing security permissions.

I'll also add @jfriend00 comment in here since I believe it's also very relevant a exposes other differences:

The biggest practical difference is that you have to design a browser application to work in an installed base of existing browsers including older versions (lowest common denominator). When deploying a node application, you get to select the ONE version of node that you want to develop for and deploy with. This allows node developers to use the latest greatest features in node which won't be available across the general browser population for years. @jfriend00

Projects like browserver are interesting and I am all for experimental development, but are they truly useful in practice? Libraries should be designed for the environment in which they are truly useful. There are no benefits in making all libraries available in both environments. Not only that it will generally result in an increased code complexity, some features will sometimes not be shimmable resulting in an inconsistent API between platforms.