Someone already mentioned TAOCP, SICP and "Introduction to Algorithms". They are all great. I'll add "Types and Programming Languages" by Benjamin Pierce to that list, or alternatively "Practical Foundations for Programming Languages" by Robert Harper. If I were you I'd bring some ebooks on Haskell (+ a cache of libs with their API docs): "Programming in Haskell" by Graham Hutton, "Real World Haskell" by by Don Stewart, Bryan O'Sullivan, and John Goerzen.

A previous comment suggests that the lifetime of most FORTRAN code is not necessarily that long (requiring scientists to publish their code might change that). If this is indeed the case, you have a point. However, most programs live very long indeed (several decades is not at all exceptional), so the productivity of the original author is not that important. If we would look at the cost over the lifetime of the code (including testing, debugging and operational cost of bugs), I think there are significant differences between languages.

Denton's opinion tell us a lot about the kind of web site he visits. The internet is a heterogenous place, and this is reflected in the quality of comments posted to various online forums, just like it is in the quality of the content posted by web site editors (CNN.com is throwing stones in glass houses posting an article like this). Look at a web site like Lambda The Ultimate. The quality of posts there is often on par with peer-reviewed journal articles. The Haskell subreddit also often has incredibly valuable discussions, all provided for free by the readers themselves.

Automatic law enforcement is cheap but it's not the way to go. Make it illegal and slap offenders with a hard punishment to deter people. It works well enough with all other reasonable laws, so why do it differently with this one?