What about performance?

Issue #85

That’s the question I hear often when people are introduced to a new framework. It’s a valid concern. But it seems to me that they ask this just for fun. To my surprise, most people just don’t care, and the frameworks with the most stars often perform the worst.

Now take a look back at performance. Here are some benchmarks

From https://github.com/ibireme/YYModel, compare different JSON mappers for ObjC

From https://github.com/bwhiteley/JSONShootout, compare different JSON mappers for Swift

From https://github.com/onmyway133/DeepDiff#among-different-frameworks, compare different diffing frameworks

I use it because it has many stars

Take a look at the stars, the ones with the most stars often perform the slowest 🙀

I don’t say that more stars mean better. I don’t believe in stars. Stars may just be a result of your marketing effort. The same framework, without any code change, but after featured in some newsletters, gets additional thousand stars. The code remains the same, so what do stars really tell here?

I’m not talking about closed source. I like open source. When deciding an open source framework, there are many factors. It can be issues and pull requests that indicate how the community care about it. It can be good code and good tests, that make it easy to maintain. It can be good documentation, that says how much dedication the developers have put in.

And here’s the fact, when you see a project with many stars, you tend to star it too 😉 for the sake of bookmarking. Stars mean little, but they give us some ideas on how popular a project is.

I just need to get work done

OK.

What about performance?

Honestly, do you really care?

Comments