Ask HN: Moving off PHP with Meteor? What are the issues?

[robomartin]

Reading-up on Meteor and related technologies I have been wondering if this concept of JavaScript everywhere might be the right way to move away from PHP. I personally like the idea of a single language owning both sides of the conversation. With frameworks like Backbone.js this seems even more interesting.

I have not done a lot of JavaScript other than the requisite level of UI stuff with JS and JQuery. I've been interested in Node.js for a while. Meteor is inspiring me to want to dive deeper into that pool and see what it can do. I have an upcoming project that was going to go the route of PHP + MySQL + JQuery. Now I am considering Meteor + MongoDB.

I'd like to hear form those with more experience and insight on the realities of making such a switch. How scalable is something like Meteor? Let's say you start hitting millions of users. What happens then? Does Meteor place a greater load on your servers in order to support all of its flexibility and "slickness"? Could something really large, like a Facebook, rely on the Meteor approach or is it better suited for small to medium projects?

I love the idea of ditching PHP. I just want to do it for the right reasons.

[manuscreationis]

Let Meteor mature until you just ship. At the very least, until the security issues are addressed.

But in the meantime, learn Node, learn Ruby, learn something else.

Don't just sit and spin because the shiniest new thing isn't perfect yet. Experiment.

[tgriesser]

If you've not done a lot of javascript, then your first priority before considering doing a framework like node.js should be to learn javascript, correctly.

Read "Javascript: The Good Parts" and then continue on to any of the recommended books/readings from there (Javascript Patterns, High Performance Javascript).

jQuery is great for DOM manipulation and XHR, but it hides a lot of the implementation details from the developer, and it'll be frustrating if you try to do anything advanced in javascript without fully appreciating the language.

So, you should be able to explain and fully comprehend:

  - Variable Scope
  - Object.prototype
  - .call, .apply, and .bind
  - Closures/Lexical Scope
  - Anonymous functions
  - Function Hoisting

Once you get down most of that, then you can consider using javascript as a primary language. Otherwise you'll be very frustrated. I'd personally recommend getting comfortable with something like jQuery + backbone.js with your current stack before diving into 100% javascript.

Also, don't worry much about scale, worry more about building something and then identifying the bottlenecks later to scale.

[robomartin → tgriesser]

I probably have dozens of languages under my belt (not keeping count) so learning a new language isn't a problem. As they say: "'been there, 'done that".

In other words, I am not the kind that will stay away from a product just because the language is new to me or I don't have a ton of experience using it. In the end, if you've been around programming for a while there really aren't very many new concepts that one couldn't pick-up quickly enough.

The first time I had to implement a genetic solver in Objective-C I was new to the language. Sure, it took a little longer to optimize it, but I knew what I had to do and it was just a matter of fitting it to the language. No big deal. At least that's how I think, valid or not.

Good books are the key. I've had "Javascript: The Good Parts" on my list for a while. Pulling the trigger today. I did go through the Sitepoint books about a year ago.

In thinking about it it seems like the most sensible bit of advice is to let Meteor mature before jumping in. I think I am going to stick with the existing game-plan until then. There will be plenty of front-end Javascript/JQuery/Backbone to write. This should pave the way to considering Meteor later on.

[edwardtrunk → tgriesser]

2nd that!!! Create a small 100% javascript client side application and don't use any libraries or frameworks, and implement using MVC approach. Going thru an exercise like this will really help out when working with the frameworks and give you a nice background with Javascript. One of the things that can be a bit confusing for folks new to Javascript is how OO and Inheritance is done. I would recommend this book: Pro Javascript Design Patterns and http://howtonode.org/object-graphs this 3 part tutorial

[joshontheweb]

Meteor is not ready for prime time as said by the creators themselves. In the mean time learn Node.js / Backone.js. Then you will be in a good place to embrace something like Meteor once it's feasible.