Hacker News Reader

Comment by Osiris | Hacker News Reader

It reminds me of Internet Explorer. Once it got "good enough", development slowed to a trickle until someone finally got fed up with it and started a new browser with a very rapid release cycle that continues today.

Node.js 0.10.0 was released on March 11, 2013. 0.12.0 was released almost 2 years later on Feb 6, 2015. By all accounts 0.12 isn't even a major change over 0.10.

The io.js team has done more in a few months than Joyent did in 2 years.

I, personally, don't care if node.js and io.js ever merge. One is clearly a superior product with a superior open source governance model.

View on HN

> The io.js team has done more in a few months than Joyent did in 2 years.

This seems a bit one-sided. What Joyent has been doing is called "maintaining production software." That's something I've been missing in the web world lately. It's surprisingly hard to find high quality libraries that don't change all the time (that is, break all the time). Libraries like jQuery, normalize.css, and Lodash are the exception here. They're some of the few libraries that I'm comfortable guessing will look pretty much the same they do now a year from now.

That is not to say that I don't want to write my gulpfiles in ES6 as much as everybody else, but you have to understand that this whole forking thing scares off people who are looking for a platform they can invest significant time and money into (that is, write more than a million lines of code for).

Why can't they have a development branch while still keeping their LTS branch?

Have they figured out how they are doing LTS yet? Last I saw there still wasn't agreement.

General consensus on hacker news seems to be angularjs making breaking changes bad, io.js making breaking changes good.

Which breaking changes are there in io.js? I'm just curious, not been following it closely; but AngularJS 2.0 is a completely different API. You can't "migrate" an Angular 1 project to Angular 2, you need to rewrite it. That's understandably scary and horrible for some people.

> Iojs: 49 authors have pushed 168 commits to master and 572 commits to all branches. On master, 8,474 files have changed and there have been 1,524,012 additions and 708,276 deletions.

With that much change in such a short time, it would be extremely optimistic to think something didnt break. If I had a lot of production code depending on node, those numbers would scare the crap out of me.

You can absolutely migrate angular 1 to angular 2. Following good practices (i.e not having 4k line controllers) that were place since its inception only makes that migration easier.

For the most part, almost all JavaScript code and modules will run the same under io.js as node.js. The primary point of breakage is binary modules. io.js is constantly upgrading their V8 bindings, so binary modules need to be updated much more frequently than with node.js.

It's more a case of dealing with the legacy problem of node.js never having upgraded V8 as often as they should have.

Server code is much easier to test with unit and integration tests than UI code which is part of the reason that people dislike breaking changes in UI frameworks more than in other kinds of libraries.

There's an interesting defence of Joyent here: https://news.ycombinator.com/item?id=9491547

Reply on news.ycombinator.com