In the past I’ve often waxed lyrical about the benefits of reading other people’s work in other disciplines and not re-inventing the wheel. But now I wonder if it’s really practical advice. Does it truly take less effort to trawl through mountains of literature than it does to simply think for a few minutes and come up with a solution yourself? The only answer I can come up with is: In some cases yes, in some cases no, but not always.
About LW, it’s true that a lot of what people have done here is simply re-invent concepts from philosophy. But in doing so they’ve given a different perspective to existing ideas, and looked at core issues rather than carrying along millenia of philosophical baggage.
A lot of mathematics, for instance, arose many times independently, and that was how people knew what stuff was important. When a concept gets reinvented many times, by people working on vastly different problems, you know it’s something that is important. Calculus, group theory, fourier analysis, etc. all fall under this definition.
This is in the class of “coordination problems”, and as we all know those are hard. I think it’s worthwhile to spend more energy on this particular one, though.
Related anecdote: one of my programmer friends uses the following algorithm for writing certain kinds of programs: “think of a plausible name for such a program had it already existed, then google for that.” This seems to work very often!
In the past I’ve often waxed lyrical about the benefits of reading other people’s work in other disciplines and not re-inventing the wheel. But now I wonder if it’s really practical advice. Does it truly take less effort to trawl through mountains of literature than it does to simply think for a few minutes and come up with a solution yourself? The only answer I can come up with is: In some cases yes, in some cases no, but not always.
About LW, it’s true that a lot of what people have done here is simply re-invent concepts from philosophy. But in doing so they’ve given a different perspective to existing ideas, and looked at core issues rather than carrying along millenia of philosophical baggage.
A lot of mathematics, for instance, arose many times independently, and that was how people knew what stuff was important. When a concept gets reinvented many times, by people working on vastly different problems, you know it’s something that is important. Calculus, group theory, fourier analysis, etc. all fall under this definition.
This is in the class of “coordination problems”, and as we all know those are hard. I think it’s worthwhile to spend more energy on this particular one, though.
Related anecdote: one of my programmer friends uses the following algorithm for writing certain kinds of programs: “think of a plausible name for such a program had it already existed, then google for that.” This seems to work very often!