I think the decision to ignore mining in the reference client, was the wrong one.
This, I wholeheartedly agree with. Two years ago core devs were saying mining had almost nothing to do with the core implementation which was a very shortsighted view. The lack of emphasis on the delicate interplay between mining and the reference implementation of the bitcoin network protocol and the block chain was a major setback. Of course one could ask why I didn't get involved then since I was busy hacking on mining code and the answer is a simple one: I don't have any faith in my c++ coding skills so it would have been presumptuous of me to try and hack on bitcoin core.
You know what?
That's the level of the argument we should be having about the big-blocks vs multi-layer debate. The problem is that bigblockers took it to the audience and now it's exclusively politics and we have a complete shitshow of a debate with people who have no idea what they are talking about.
If it was the level set by people like dinofelis who actually have very respectable points, then this would be MUCH better. The problem is that the bigblocker side has degenerated to the point their position is untenable. I think this didn't start completely by a problem of their own, but from a separate problem which is that of software development governance. But the reality is that every client they come with is worse than the previous one and we cannot abstract the underlying technical debate from the fact that their current dev base is completely incompetent as 95%+ of devs with experience in this field (which is a HARD field) "roughly" agree with the Core Roadmap or are willing to make concessions to work with them (me included, although I stay very anonymous in dev because this could cost me my job in finance).
So I feel we don't have a way out that doesn't end up in some sort of radicalisation and infighting. They will resort to aggression, losing the argument on that alone even though they do have fair points like those exposed by dinofelis (I just disagree in his conclusions and in some key assumptions, but I actually agree with a lot of his take).
I'm seriously worried about this. If Bitcoin depends on sorting the problem of "perfect agreement in software development" then we are IN THE SHIT because that is never going to happen.