4 ms·

I somewhat doubt your assessment. Aside from the fact that the article references parts of the paper, my impression of the author is that he's the kind of perso

I somewhat doubt your assessment. Aside from the fact that the article references parts of the paper, my impression of the author is that he's the kind of person who might have actually read the paper. C.f. http://conal.net/cv.pdf http://conal.net/cv.pdf

imode 8y ago
"Read" is not "skimmed". My assessment is due to the fact that functional programming was not what Backus had in mind, and FP reflects that. FP reflects an algebra of programs, not traditional functional programming. It doesn't suffer from these problems that the author is stating. If the author read said paper and not skimmed it, they might have been on the right track. EDIT: Downvoting me because you disagree with me is petty, childish, and really does say a lot about the user base of this site.
- hyperpape 8y ago
  As far as I can see, Conal does not state that FP has problems. As I read it, he sees a promise in Backus’s lecture that is not realized by contemporary functional programming.
  - imode 8y ago
    His mistake is thinking that contemporary functional programming can actually realize Backus' promises. Backus' method of programming (FP) is notably different than functional programming, and as such admits itself to a different style of reasoning. The author wouldn't need to write this article if they simply adhered to Backus' FP, and not used contemporary functional programming. This site simply doesn't allow discourse anymore. Not unless you already agree with somebody. Shame.
    - sifoobar 8y ago
      Forgive them, for they know not what they do. The people running this site though. The whole thing is designed to make life very difficult for anyone who steps out of line. And yes, it does seem to get worse and worse from here.
    - rocqua 8y ago
      > This site simply doesn't allow discourse anymore. Not unless you already agree with somebody. Shame. Hi, I really like discourse. Especially with people who have strong opinions in a field I'm learning about. Notably, 'functional programming' is such a field, and you seem like such a person. However, statements like the above kind of turn me off from discourse. It heavily suggests there will be friction in the conversation. Thus, I would ask you to leave out the comments. Your points will still be made, and made more concisely. Moreover, it improves the chance someone will actually reply. This is doubly so when responding to downvotes, rather than responding to negative replies. Because downvotes don't look nearly as aggressive as negative replies. Moreover, the person reading your response is probably not the one who downvoted.
    - sctb 8y ago
      You could consider investigating why your comments are so consistently flagged on this website. We're here to read informative things, not snark about what so-and-so knows or does not know. So please just teach us about Backus' FP instead of going on an information-free tirade.
- carapace 8y ago
  (Stop commenting on the downvotes dude, people will pile on for it. Also, who cares? "Ignore the bird, follow the river.") > FP reflects an algebra of programs, not traditional functional programming. It doesn't suffer from these problems that the author is stating. You're so right. (Not about Conal Elliot not having grokked the paper though, see below.) The Joy programming language is much closer to Backus' FP than Haskell, for example, and working with it is much more like deriving mathematical equations than any other language I've tried (FWIW.) (Some examples here: http://joypy.osdn.io/notebooks/index.html http://joypy.osdn.io/notebooks/index.html) He's very much on the right track though! http://conal.net/papers/compiling-to-categories/ http://conal.net/papers/compiling-to-categories/ Here he's converting Haskell code to a point-free style that is very similar to Joy, then instantiating the code over different Categories to get multiple actual programs/outputs from a single expression: "a variety of interpretations including hardware circuits, automatic differentiation, incremental computation, and interval analysis."
  - imode 8y ago
    Now these are some cool resources! Thank you for providing them. I'm really starting to enjoy Joy, J, FP, and other languages based on a similar combinator-like basis. I'll have to read over Compiling to Categories, because that certainly is something that catches my eye.
    - carapace 8y ago
      Cheers! I have a hunch that the Categorical programming paradigm will prove to be really cool. I hope it doesn't take decades to get into common use, eh? You might also like this version of Joy implemented in Prolog: https://osdn.net/projects/joypy/scm/hg/Joypy/blobs/tip/thun/thun.pl https://osdn.net/projects/joypy/scm/hg/Joypy/blobs/tip/thun/... Check out the "compiler" at the bottom.

imode 8y ago

"Read" is not "skimmed". My assessment is due to the fact that functional programming was not what Backus had in mind, and FP reflects that. FP reflects an algebra of programs, not traditional functional programming. It doesn't suffer from these problems that the author is stating. If the author read said paper and not skimmed it, they might have been on the right track. EDIT: Downvoting me because you disagree with me is petty, childish, and really does say a lot about the user base of this site.

hyperpape 8y ago
As far as I can see, Conal does not state that FP has problems. As I read it, he sees a promise in Backus’s lecture that is not realized by contemporary functional programming.
- imode 8y ago
  His mistake is thinking that contemporary functional programming can actually realize Backus' promises. Backus' method of programming (FP) is notably different than functional programming, and as such admits itself to a different style of reasoning. The author wouldn't need to write this article if they simply adhered to Backus' FP, and not used contemporary functional programming. This site simply doesn't allow discourse anymore. Not unless you already agree with somebody. Shame.
  - sifoobar 8y ago
    Forgive them, for they know not what they do. The people running this site though. The whole thing is designed to make life very difficult for anyone who steps out of line. And yes, it does seem to get worse and worse from here.
  - rocqua 8y ago
    > This site simply doesn't allow discourse anymore. Not unless you already agree with somebody. Shame. Hi, I really like discourse. Especially with people who have strong opinions in a field I'm learning about. Notably, 'functional programming' is such a field, and you seem like such a person. However, statements like the above kind of turn me off from discourse. It heavily suggests there will be friction in the conversation. Thus, I would ask you to leave out the comments. Your points will still be made, and made more concisely. Moreover, it improves the chance someone will actually reply. This is doubly so when responding to downvotes, rather than responding to negative replies. Because downvotes don't look nearly as aggressive as negative replies. Moreover, the person reading your response is probably not the one who downvoted.
  - sctb 8y ago
    You could consider investigating why your comments are so consistently flagged on this website. We're here to read informative things, not snark about what so-and-so knows or does not know. So please just teach us about Backus' FP instead of going on an information-free tirade.
carapace 8y ago
(Stop commenting on the downvotes dude, people will pile on for it. Also, who cares? "Ignore the bird, follow the river.") > FP reflects an algebra of programs, not traditional functional programming. It doesn't suffer from these problems that the author is stating. You're so right. (Not about Conal Elliot not having grokked the paper though, see below.) The Joy programming language is much closer to Backus' FP than Haskell, for example, and working with it is much more like deriving mathematical equations than any other language I've tried (FWIW.) (Some examples here: http://joypy.osdn.io/notebooks/index.html http://joypy.osdn.io/notebooks/index.html) He's very much on the right track though! http://conal.net/papers/compiling-to-categories/ http://conal.net/papers/compiling-to-categories/ Here he's converting Haskell code to a point-free style that is very similar to Joy, then instantiating the code over different Categories to get multiple actual programs/outputs from a single expression: "a variety of interpretations including hardware circuits, automatic differentiation, incremental computation, and interval analysis."
- imode 8y ago
  Now these are some cool resources! Thank you for providing them. I'm really starting to enjoy Joy, J, FP, and other languages based on a similar combinator-like basis. I'll have to read over Compiling to Categories, because that certainly is something that catches my eye.
  - carapace 8y ago
    Cheers! I have a hunch that the Categorical programming paradigm will prove to be really cool. I hope it doesn't take decades to get into common use, eh? You might also like this version of Joy implemented in Prolog: https://osdn.net/projects/joypy/scm/hg/Joypy/blobs/tip/thun/thun.pl https://osdn.net/projects/joypy/scm/hg/Joypy/blobs/tip/thun/... Check out the "compiler" at the bottom.