I had some very good and also very bad experiences and choices with open-source and would love to know some of yours!
What are the key aspects you look for when choosing an open source library or framework?
I had some very good and also very bad experiences and choices with open-source and would love to know some of yours!
What are the key aspects you look for when choosing an open source library or framework?
For further actions, you may consider blocking this person and/or reporting abuse
alexandre-emmanuel -
Jai Bhullar -
thinkThroo -
Vitor sabino -
Top comments (6)
Thereโs this good blog post by automation panda which speaks about all-in-one vs piece-by-piece frameworks I found recently. Sometimes you just want a quick and dirty big library that gets the job done, which is great if thereโs awesome documentation and a strong active developer community backing it (bug fixes over just bug reports). However, it can feel like your code is locked in to that one bubble of a framework. There are some fantastic big projects suffocating under a backlog of bugs, and itโs sad to see, but thatโs usually when you need to look for an alternative framework.
I find the piece by piece approach better, maybe because I like the UNIX philosophy. Still, for one thing I want to do, I try to find at least two frameworks that can do the job, and have that in mind when I write my code. If I choose the bleeding edge little library and the developer goes missing in action, I could switch to the stable big library. If the big library gets too buggy, I switch to the nimble one. Thatโs the beauty of open source really, but it does mean a lot of work. All depends on how long term your code is meant to be.
Cool article, agree mostly, it's a good categorization.
Also from a learning perspective, if your going to work on a codebase for a long time you'll need to know what happens under the hood with the given framework. Even though you need abstraction, and don't wan't to get into the frameworks implementation.
I remember when I started developing a project in Rails for my first client, I knew the basics about Ruby and Rails, at the beginning it didn't matter, the incredible setup experience, generators and meta-programming magic let me move forward rapidly. But when time passed by and I started needing to implement more complex features, and also errors began to appear, I happened to be a bit lost (after that I read Metaprogramming in Ruby - great read by the way - in some days and fell in love with the language and Rails' magic modules like Active Record).
Is not that is the framework's fault, Rails is a great framework, but for beginners, all-in-one solutions can be a double-edge sword sometimes, with some years of experience I'm also more inclined to piece-by-piece solutions (Unix commend style, node modules, microframeworks, etc).
Documentation, ease of use and community size mainly.
I don't really care if there's a big company behind it or not
Thanks for the feedback rhymes! Did you ever regret having chosen a library?
Who didn't :D
The good thing is that, if you layer your code correctly, you should be able to replace such library without too much effort.
Just wrap it in a sort of facade so that you don't call
library.method()
everywhere in your app, just in a limited amount of places (ideally one or two files).This way, if you need to swap it with a better one, the damage is limited
Agree 100%, good design and abstraction always reduces risk