[markshuttle]

There are always lots of opinions about how to build a piece of software. Unlike a hundred yard dash, where the best performance is obvious to all, the risks and benefits of different engineering approaches remain a mystery until years after some work is done.

I've been around for those years :)

When we need something difficult and fundamental thought through very deeply, because it has to be precisely correct, I've learned to trust our CTO. Others usually want to go faster, or fix-it-later, or we-don't-need-a-spec-just-be-agile. That creates tension. But when I think it needs to be done very carefully and correctly, some tension is worthwhile.

I wish we had been more careful with some of the foundations for Unity, for example, because they would then have served others better even if the project itself failed.