3 ms·

OP makes surprising claims about J. Here is some reading which hopefully will help you make up your own mind about whether the claims are accurate. J has a ric

by jph00 8y ago

OP makes surprising claims about J. Here is some reading which hopefully will help you make up your own mind about whether the claims are accurate. J has a rich range of composition operations (far more than mentioned in OP) which work well together. Many are discussed in this paper: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.98.3105&rep=rep1&type=pdf http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.98.... Matrix convolution is handled with semidot3: https://code.jsoftware.com/wiki/Vocabulary/semidot3 https://code.jsoftware.com/wiki/Vocabulary/semidot3 FFT convolution examples are here: https://code.jsoftware.com/wiki/Essays/FFT https://code.jsoftware.com/wiki/Essays/FFT Iverson wrote a chapter on polynomials in J in his book Exploring Math: http://www.jsoftware.com/books/pdf/expmath.pdf http://www.jsoftware.com/books/pdf/expmath.pdf

jonahx 8y ago
Indeed. J is a wonderful language. From the article: > I don't know why Iverson thought the hook was the thing to embed in the language. There are deep reasons for the choice: http://www.jsoftware.com/papers/fork1.htm http://www.jsoftware.com/papers/fork1.htm In addition, viewing J as "a worse Haskell" is to misunderstand it. J is a function-level, not a purely functional, language: https://en.wikipedia.org/wiki/Function-level_programming#Contrast_to_functional_programming https://en.wikipedia.org/wiki/Function-level_programming#Con... From that link: > When Backus studied and publicized his function-level style of programming, his message was mostly misunderstood,[2] as supporting the traditional functional programming style languages instead of his own FP and its successor FL. > ... > This restriction means that functions in FP [Backus's function-level language] are a module (generated by the built-in functions) over the algebra of functional forms, and are thus algebraically tractable. For instance, the general question of equality of two functions is equivalent to the halting problem, and is undecidable, but equality of two functions in FP is just equality in the algebra, and thus (Backus imagines) easier.

by jph00 8y ago

jonahx 8y ago

Indeed. J is a wonderful language. From the article: > I don't know why Iverson thought the hook was the thing to embed in the language. There are deep reasons for the choice: http://www.jsoftware.com/papers/fork1.htm http://www.jsoftware.com/papers/fork1.htm In addition, viewing J as "a worse Haskell" is to misunderstand it. J is a function-level, not a purely functional, language: https://en.wikipedia.org/wiki/Function-level_programming#Contrast_to_functional_programming https://en.wikipedia.org/wiki/Function-level_programming#Con... From that link: > When Backus studied and publicized his function-level style of programming, his message was mostly misunderstood,[2] as supporting the traditional functional programming style languages instead of his own FP and its successor FL. > ... > This restriction means that functions in FP [Backus's function-level language] are a module (generated by the built-in functions) over the algebra of functional forms, and are thus algebraically tractable. For instance, the general question of equality of two functions is equivalent to the halting problem, and is undecidable, but equality of two functions in FP is just equality in the algebra, and thus (Backus imagines) easier.