3 ms·

It is a tool of thought _for smart people_. But thanks for that, um, like, uh totally awesome Wikipedia link! +1

civilized 5y ago
> It is a tool of thought _for smart people_. As I said, if this were really true, there would be evidence of people using it to express thoughts that are too hard to express by other means. But in reality, people who think in APL just don't seem to be on the forefront of published algorithmic thought. It might be helpful to contrast APL with nonstandard analysis, a mathematical tool of thought for doing rigorous real analysis more easily. People have proved new and useful results in nonstandard analysis, and the fact that these results were first proved in nonstandard analysis is at least some evidence that nonstandard analysis served as a real tool of thought to produce those results. (To be fair, the idea that nonstandard analysis was an essentially tool in these proofs is controversial among mathematicians. But there is at least the hard fact of the results being proved first by this means.) So, if smart people are able to use APL to push the boundaries of algorithmic thought, we'd expect APL-ers to have that sort of achievement to show for it. But we don't see that. After pre-existing thoughts are translated into APL's logography, there does not seem to be any demonstrable enhancement of the ability to reason about algorithms. Or maybe there is some enhancement for that individual person, relative to their abilities without APL, but it does not seem to cause that person to rise above the general human level of algorithmic ability. "APL as a tool of thought" seems to be a beautiful theory whose predictions have not met with reality.
- patrec 5y ago
  > As I said, if this were really true, there would be evidence of people using it to express thoughts that are too hard to express by other means. Like Arthur Whitney you mean?
- dzaima 5y ago
  What sort of evidence are you expecting there to be? A tool of thought would usually, uh, stay as thought. I'd expect a good percentage of APL users are using it as such, but making a big deal out of that just doesn't really make any sense for any given individual. At least in the couple array language chat rooms I'm in, pseudocode-ish APL is somewhat frequently used to convey thoughts, and I certainly use arraylangs as such too. APL doesn't magically double your ability to think though. Pushing boundaries takes a lot more than raw thought power. And APL is used so little that you couldn't make any correlations even if APL increased thinking ability tenfold anyways. But it still is a very useful tool (among many other tools) from time to time. I personally don't see myself using any array language as a general purpose language for everything though. They're useful (extremely useful!) for things that map well to array operations or benefit from terseness, and anything from "meh" to awful for other things.

> It is a tool of thought _for smart people_. As I said, if this were really true, there would be evidence of people using it to express thoughts that are too hard to express by other means. But in reality, people who think in APL just don't seem to be on the forefront of published algorithmic thought. It might be helpful to contrast APL with nonstandard analysis, a mathematical tool of thought for doing rigorous real analysis more easily. People have proved new and useful results in nonstandard analysis, and the fact that these results were first proved in nonstandard analysis is at least some evidence that nonstandard analysis served as a real tool of thought to produce those results. (To be fair, the idea that nonstandard analysis was an essentially tool in these proofs is controversial among mathematicians. But there is at least the hard fact of the results being proved first by this means.) So, if smart people are able to use APL to push the boundaries of algorithmic thought, we'd expect APL-ers to have that sort of achievement to show for it. But we don't see that. After pre-existing thoughts are translated into APL's logography, there does not seem to be any demonstrable enhancement of the ability to reason about algorithms. Or maybe there is some enhancement for that individual person, relative to their abilities without APL, but it does not seem to cause that person to rise above the general human level of algorithmic ability. "APL as a tool of thought" seems to be a beautiful theory whose predictions have not met with reality.

patrec 5y ago
> As I said, if this were really true, there would be evidence of people using it to express thoughts that are too hard to express by other means. Like Arthur Whitney you mean?
dzaima 5y ago
What sort of evidence are you expecting there to be? A tool of thought would usually, uh, stay as thought. I'd expect a good percentage of APL users are using it as such, but making a big deal out of that just doesn't really make any sense for any given individual. At least in the couple array language chat rooms I'm in, pseudocode-ish APL is somewhat frequently used to convey thoughts, and I certainly use arraylangs as such too. APL doesn't magically double your ability to think though. Pushing boundaries takes a lot more than raw thought power. And APL is used so little that you couldn't make any correlations even if APL increased thinking ability tenfold anyways. But it still is a very useful tool (among many other tools) from time to time. I personally don't see myself using any array language as a general purpose language for everything though. They're useful (extremely useful!) for things that map well to array operations or benefit from terseness, and anything from "meh" to awful for other things.