I agree with you entirely. A huge problem is in mindset and culture. I've mentioned formal methods here on HN before and been essentially shouted down. Which seems odd considering the hacker-above-the-commoners ethos HN wears on its sleeve.

But I digress, I think in the next few years as functional languages become more mainstream we'll see a huge uptake in formal methods. Not end to end mind you, but of the low level proover type tools.

Formal project planning and design languages still take too much training to be widely accepted. Can you imaging a ruby hacker analyzing and writing up his code in Z language before they code? Neither can I. But any amount of progress on this front is a good thing.