>>12392598Spivak is a million times better than Stewart. For mathematicians, at least. I suppose engineers would prefer the latter.
>>12389914I would put How to Solve it at the very beginning, and it would be obligatory. Does far more than 3 whole books on logic will.
Not to say I'd leave logic out of this, but I don't know books on logic to suggest a good one. Enderton's Elements of Set Theory, perhaps?