I’ve been working with Cocos since 2015 when I started to work for the company that I work until now. I started with cocos2d-js + Cocos Studio. It was a little complicated to export the scene in one software, and code/build in other, but it did the trick. The js -> cpp interpretation (then the build for java/objective-C) was really annoying and gave me a lot of headache.
I don’t know how much time I worked with 2d-js, but then I went to work with Unity for something between 1~2 years.
When I came back to Cocos, back in 2017, it was already on CocosCreator 1.6.1. The jump from 2d-js to CC was really good, finally an visual editor that I could drag the code for the nodes, had scene structure, could code in JS, yay!
I really enjoyed the CC, almost became an evangelist in my college, always talking about it with the people interested in games, gave some workshops, and so on. I really hoped that cocos community grew up, so we could help each other. But it didn’t.
The hard part was not even convincing to use it over Unity, it was that it will work as designed. The most common complaint was that the documentation said that this element will work some way, but it didn’t. And it was not only in my college, every new teammate that started a new project on Creator had the same complaint. The amount of workarounds that we talked to my teammates, was like, as we say in Brazil, we were stepping in eggs, always avoiding some approaches because it will break that way.
I’m using the 1.9 version from the week that it launched, but the iOS change in its libraries made we port the game for some more recent version. First I went to 2.3.3, then when 2.4 launched was a dream coming true because it will solve a lot of problems with distribution over our projects.
And now I am facing this lack of reliability, but without that joy that I had when I met CC, neither with the hope that some day it will be better. What looks like to me is that the priority is to implement as much feature as possible in the shortest amount of time, but not caring at all if something will break in the way and/or what is being released is really working.
OK, I get that fast releases is a good thing, CC is way behind other engines in the number of features and this is something to run as fast as it can, we are all software developers after all, but hell, even that, we don’t have an schedule of the next releases with was intended for that version! And the patch releases only with adding a new build target really doesn’t help!
I’ll continue to work with CC because my work needs to, keeping my eye on where it is going and trying to make my best to get as much as it can offer. But today, if someone ask me what engine I would recommend, Cocos is not one of them.