The code we "want" to write is that which matches the way we think .... [snip] My way is to think hard about what the best way to think about things is. I'm in two minds. On the one hand, we're in violent agreement: The code we /want/ to write is that which matches the way we /want/ to think, genuflecting, as it were, before the cold altar of mathematical perfection. On the other hand, with a view towards AI, I'd want to code just the way I now currently think, warts and all, except that I've designed the DSL to fix all my idiosyncrasies, hidden contexts, annoying ambiguities, utter silliness, etc. You might claim that the former is easier and more achievable. I don't see why. It seems likely that the kind of perfection you seek can only be obtained by piercing insights into the nature of the mind's present drosses (if indeed they be so). If you do reach such lofty heights of re-cognition, why not just program all of that into the DSL so that it corrects for all those mental lapses? Surely no more difficult than retooling the mind permanently. Plus you get to share the wealth: make it easy for others to reach programming nirvana too. (We're already there: it's that tarpit called Perl *smacks forehead*) The completed journey ends with a return to where one first began. Or something like that. mathematicians spend a lot of time thinking about how to think about things. G.C. Rota (RIP): The philosopher's role has always been that of stating facts that may have been on everybody's mind but that no one dared state clearly. -- View this message in context: http://old.nabble.com/Finally-tagless-and-abstract-relational-Algebra-tp2694... Sent from the Haskell - Haskell-Cafe mailing list archive at Nabble.com.