I have no issues with his engine or anything. My issue is when people complain about others' work without understanding how it actually works. Of course LibGDX is slower. Saying your engine is better because it is faster (thus insulting someone else's work) is, at least in my book, ignorance and it is very annoying.
I agree with that, really a public benchmark should be provided with any claims of being 'better' than something else. It's quite possible SwordsMiner used a non-optimal method in whatever he is basing his libGDX performance baseline on. A benchmark and results would also tell as to just how much of an improvement there actually is. (it could be negligible)
- is fast.
- gets the job done.
libGDX "gets the job done" much faster (and likely more thoroughly) than rolling your own solution, in terms of development time in the vast majority of cases, and is plenty fast for practical applications. That said, the NIH principle facilitates learning many underlying concepts through experience, so carry on. Just realize that that is it's benefit, not the ones you are specifying.