Answer 1

I'd argue that creating your own will be so complex, time consuming and full of bugs, that it wouldn't be feasible. I'd always try to find an engine that best fits what you want.

Answer 2

I Agreed, Building your own is good even the result is bad and it can increase your experience about programming. If you dont want to make your own maybe you can try UE4 i guess it support real time lighting. But the engines was build with c++ and for me, if i want make some game, i will disable realtime lightning for getting a good perfomance (My laptop is old ^_^)

Answer 3

Adding on my thinking,
Creating your own engine does allow for customization as you pointed out, but it will be buggy and not properly optimized as Xan stated. You will very likely not build anything better than the top engines, they have huge support with top programmers and tools.

Building your own also helps you understand how engines work though, so i do recommend beginners building a small, not too complex, not too time wasting, 2d or 3d engine if they want to be game programmers ;)

Answer 4

Xan Agreed, never really use your engine unless its top notch for some reason xD

Answer 5

「ＨＡＰＰＹ ＴＯ ＨＥＬＰ」 Agree with both your points. Similar to web frameworks... creating your own is educational, but nobody should sensibly use their own, for a production system. Unless there is a rare good reason.