Just to add to the analogies I drew in "The Power of the Context". In baseball, an "error" (or "failiure") is not catching a catchable flyball, etc., But, striking out is not an "error" or "failure" -- it is better considered as "the overhead for sometimes accomplishing something really difficult".
To bring this analogy to Xerox Parc, the technical aspects of computing -- such as building a physical computer or an operating system or new programming language -- should be successful "98%" or more of the time (this is roughly the fielding percentage in baseball). These require technical skills for which quite a bit of knowledge and practice has to be acquired and done in advance. It is a kind of engineering, often with some new design elements, but where "the bridge has to stay up".
The really difficult parts of computing lie in attempts at inventing ways to create enormous new leverages via new kinds of organizations and designs. Here we should be ecstatically happy if we achieve the .406 that Ted Williams reached in 1941. The 60% this doesn't work out is just "overhead".
The downside of skill and knowledge is the temptation to use memory before thinking things through. The downside of ignorance and cleverness is that usually worse than "reinventing the flat tire" (which is all too common these days).
I've advocated "learning everything" and then "forgetting it except for the perfume". In other words, though "most ideas are mediocre down to bad", one has to have them freely rather than just applying technique.
The abundance of bad ideas can interfere, so you have to get rid of them somewhere. A good idea will have something like an odor to it that will allow one to find relevant knowledge in the past (often a very different past than the one that led to the present). This knowledge will help vet the idea, and eventually allow the weakest part of the process -- one's cleverness -- to possibly do something worthwhile for once.
Just to add to the analogies I drew in "The Power of the Context". In baseball, an "error" (or "failiure") is not catching a catchable flyball, etc., But, striking out is not an "error" or "failure" -- it is better considered as "the overhead for sometimes accomplishing something really difficult".
To bring this analogy to Xerox Parc, the technical aspects of computing -- such as building a physical computer or an operating system or new programming language -- should be successful "98%" or more of the time (this is roughly the fielding percentage in baseball). These require technical skills for which quite a bit of knowledge and practice has to be acquired and done in advance. It is a kind of engineering, often with some new design elements, but where "the bridge has to stay up".
The really difficult parts of computing lie in attempts at inventing ways to create enormous new leverages via new kinds of organizations and designs. Here we should be ecstatically happy if we achieve the .406 that Ted Williams reached in 1941. The 60% this doesn't work out is just "overhead".
The downside of skill and knowledge is the temptation to use memory before thinking things through. The downside of ignorance and cleverness is that usually worse than "reinventing the flat tire" (which is all too common these days).
I've advocated "learning everything" and then "forgetting it except for the perfume". In other words, though "most ideas are mediocre down to bad", one has to have them freely rather than just applying technique.
The abundance of bad ideas can interfere, so you have to get rid of them somewhere. A good idea will have something like an odor to it that will allow one to find relevant knowledge in the past (often a very different past than the one that led to the present). This knowledge will help vet the idea, and eventually allow the weakest part of the process -- one's cleverness -- to possibly do something worthwhile for once.