HN Explorer

Ask HN: pros and cons of different web stacks/frameworks

5 points by choochootrain ↗ HN
I'm just starting to get into web development, and it looks like theres alot of choices out there. I dont really have an idea in mind, but I just want to start tinkering. I've heard of the LAMP stack, but I'm leaning more towards Ruby and Rails.

Some of the frameworks I've read about and considered are Rails, Sinatra, CakePHP, node.js?, and some others in Erlang, Java, and Lisp.

To me, it looks like all these frameworks provide the same functionality, but with different languages, databases, and web servers. Is there some kind of way to choose, or do you just pick one that youre familiar with?

6 comments

[ 3.1 ms ] story [ 22.6 ms ] thread
megamark16
I'm partial to Python and Django, but in the end I think it comes down to preference, and what you feel most comfortable with. Honestly, if it were me I'd find and follow a tutorial for creating a small web app (like a blog) in a few different frameworks and see which one feels right for you.

There was a discussion a few years ago on HN about this very topic, although it was mostly focused on Django VS Ruby on Rails, but there is lots of good discussion in the thread all the same. Here's the link:

http://news.ycombinator.com/item?id=195423

Welcome to the web dev community! Come visit us on the Django IRC channel (webchat.freenode.net, channel #django) if you have any questions about django.

mindcrime
To me, it looks like all these frameworks provide the same functionality, but with different languages, databases, and web servers.

Basically, yeah.

Is there some kind of way to choose, or do you just pick one that youre familiar with?

To be honest? There are too many out there, to actually do a thorough analysis / review of each, on your own. It'd take you 10 years, and by the time you got done, it would all be obsolete. So, ultimately, you have to fall back to some kind of heuristic: recommendations from trusted friends / colleagues / etc., simple popularity, some "back of the envelope" calculations of some meaningful metric (availability of developers for that platform, for example), the aforementioned point of "go with what you know," size and activity level of the community around each one, ease of learning, etc., etc.

All of that said, I'm a big fan of Groovy on Grails, which I find has a lot going for it. But there are far more frameworks I haven't used, than ones I have, so I can't even begin to make a claim that Grails is objectively "the best" in any sense. In my case, I picked it because I already knew Java, was familiar with the JVM ecosystem, libraries, app servers, messaging systems, etc., and because Groovy is very easy to learn if you know Java. The pros I see for Grails:

1. Open source (most of them are these days, but still)

2. Easy to learn (this is relative, since I'm coming from Java)

3. Rapid development using convention of configuration (same as Rails, etc.)

4. allows me to leverage a lot of existing knowledge of using Java app servers, tools, and infrastructure.

5. Easy use of tons of useful java libraries that already exist. I can easily incorporate Tika, Lucene, Mahout, HttpClient, Commons Math, etc., etc., etc.

Cons to Grails:

1. There is some question as to the future of Groovy, since the creator of the language said something along the lines of "If I'd known about Scala when I created Groovy, I wouldn't have created Groovy." That said, the Groovy and Grails community seem to be very active, and I don't see any particular reason to think either will be dying off anytime soon.

2. Performance: Groovy has a reputation for being a little bit of a performance dog, compared to some of the other JVM based languages. But it has gotten a lot better with recent releases, and there's reason to think that the JDK7 changes (MethodHandles API, invokedynamic opcode) will make Groovy even faster.

So there ya go, a few thoughts and comments on Grails. I'd highly recommend at least giving it a look, especially if you know Java already.

choochootrain
I actually have a couple years of experience with Java. I'm looking into Ruby and Python not only because of their popular frameworks, but also because I feel like I need to broaden my horizons a bit. Do you think its worth learning Ruby/Python, or does Java/Groovy get the job done?
mindcrime
I don't really think there's anything you can do in Ruby or Python, that you fundamentally can't do in Groovy, although certain things might be easier. But I do encourage everyone to broaden their horizons, so learning Python and/or Ruby is definitely not a bad idea. Of the two, I've used Python more than Ruby, and am fairly fond of it, as non-JVM languages go. Python and Pylons (well, it used to be called Pylons, I think it's something else now) is certainly a fine stack for web apps.
brandoncordell
I think the best way to choose is narrowing it down to a few potential frameworks/stacks that you (and your team, if applicable) thinks would be the best fit for your project. I don't believe any of these frameworks are the one answer to all your problems.

At work we build an enterprise level government used application in CakePHP. It's done fairly well, but given the ability to migrate our platform I would jump at the chance.

We begun trying to practice TDD on the project and it's been nothing but a nightmare! Whether it's Simpletest or CakePHP's implementation of Simpletest we haven't figured out (or had the time to dive in and figure it out). We run into issue after issue. The littlest change to the way our app runs at the core seems to trigger a shit storm of little issues here and there. Granted, a lot of this has to do with the architecture that was put in place by the Sr. developer before I got here, but it's definitely becoming an issue...

Might also be worth noting that CakePHP is a (poor in my opinion) port of Rails to PHP.