Very helpful, thanks. And for what it's worth, this is what I (and xyzzyz above, I imagine) wished you would have said initially, two comments up, for a higher HN S/N.
Yes, helpful, but P/E is still important as a proxy measure for all of the things you mentioned. Current earnings are not a worse predictor of future earnings than many other more complex possible models. Under the assumption that future earnings will be at least reasonably predicted by current earnings, comparing the price as a multiple to that predictor is not a bad way to make a purchase decision relative to other assets. Of course book value is important, as are growth prospects, but book value (hopefully less so after IFRS4) is not really a great estimator of fire-sale value, and I would argue that growth is second order to earnings' first order and accordingly harder to estimate.