Your last argument, I find it irrelevant. Buttons and links etc. change colours not because of their inability to indicate their function, but to enhance the interactive experience. In most cases you get exactly the same effect with so-called "skeyomorphic" elements, no difference at all. The whole discussion has nothing to do with your "underlined links" argument.
The flat look has received a lot of criticism in the past couple years, but to be completely honest, in the many arguments I've heard against it, I couldn't really find one that can really stand it's ground if I may say so. On the contrary, I find a lot of merit in the new principles that are being embraced trough this approach, such as e.g. the emphasis on interaction to distinguish elements, instead of artificially imposed "symbolisms" (whether this is an underline, or a button-shaped, well, button :) ). This gives me a hint that, as a design community, we have a more mature approach to designing for those unearthly things we call "devices" than we did a few years back :)
Dear HN users: If you are even minimally interested in the topics that this post "covers", please, do yourselves a favour and open up any book about machine intelligence instead of reading such uninformed and negligent posts.
Very good point! It's not the reasoning and logic behind regex that would turn a beginner away from it, its the syntax that is being used that often results into expressions turning into impenetrable goo!
(thanks for the verbal expressions link btw, seems really promising)
Looking this up on Wikipedia reveals it's relationship with the Three Prisoners Problem[1] as well. The solution presented in that case helped me in reasoning about it.
