Re: Is Clojure's lack of support for forward referencing a feature or a bug?
On Wed, Mar 11, 2009 at 5:18 PM, Timothy Pratley timothyprat...@gmail.com wrote: It is also quite trivial to patch the compiler to auto-def symbols as it finds them instead of throwing an error. I would be interested in knowing how to do such a patch. When I work on code, I like to organize my functions in a way that makes it easy to read and understand what is going on. As I work on longer chunks of Clojure code, I'm finding that shuffling around the functions to avoid a lot of forward declarations is destroying the readability of my code. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Clojure group. To post to this group, send email to clojure@googlegroups.com To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en -~--~~~~--~~--~--~---
Re: Is Clojure's lack of support for forward referencing a feature or a bug?
Hi Mark, A fuller discussion can be found here: http://groups.google.com/group/clojure/browse_thread/thread/a99b420d5ee0aa40/47f8c2ab6845e9ae Which has links to the simple patch I tried, and discusses the more advanced technique Laurent experimented with. Elena subsequently developed an emacs plugin which looks interesting (I'm a VI ninja though so haven't used it) http://groups.google.com/group/clojure/browse_thread/thread/ca7076f4c6591fdd/cda5cf10b89a3679 My own experience FWIW was that it was great for two weeks coding with autodef, then for about a week I became frustrated with my typos and disabled it. More promising solutions might come from an external tool (such as Knuth's literate programming noweb) or IDE support like Elena described. For now my work flow is write the code backwards (ie: manually move the cursor up) and/or chopping and pasting. Then when I'm happy with it, I re-chop it all in my 'preferred' order and put a declare at the top. That sounds quite inefficient, but VI is really great for re- organizing text blocks so it is not too strenuous. That said, I'm really interested in ways that literate programming style can be followed with the least external support. Regards, Tim. On Mar 26, 4:15 pm, Mark Engelberg mark.engelb...@gmail.com wrote: On Wed, Mar 11, 2009 at 5:18 PM, Timothy Pratley timothyprat...@gmail.com wrote: It is also quite trivial to patch the compiler to auto-def symbols as it finds them instead of throwing an error. I would be interested in knowing how to do such a patch. When I work on code, I like to organize my functions in a way that makes it easy to read and understand what is going on. As I work on longer chunks of Clojure code, I'm finding that shuffling around the functions to avoid a lot of forward declarations is destroying the readability of my code. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Clojure group. To post to this group, send email to clojure@googlegroups.com To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en -~--~~~~--~~--~--~---
Re: Is Clojure's lack of support for forward referencing a feature or a bug?
For what it's worth I'm a big fan of the wishful thinking programming style. I write some functions how I think they should look, declare the functions I haven't defined yet. Then I implement the lower level functions, write some tests- then usually the higher level stuff works without too much tweaking. In a non-functional programming language this doesn't work so well, but it's a good fit for something like Clojure. On Thu, Mar 26, 2009 at 1:42 AM, Timothy Pratley timothyprat...@gmail.comwrote: Hi Mark, A fuller discussion can be found here: http://groups.google.com/group/clojure/browse_thread/thread/a99b420d5ee0aa40/47f8c2ab6845e9ae Which has links to the simple patch I tried, and discusses the more advanced technique Laurent experimented with. Elena subsequently developed an emacs plugin which looks interesting (I'm a VI ninja though so haven't used it) http://groups.google.com/group/clojure/browse_thread/thread/ca7076f4c6591fdd/cda5cf10b89a3679 My own experience FWIW was that it was great for two weeks coding with autodef, then for about a week I became frustrated with my typos and disabled it. More promising solutions might come from an external tool (such as Knuth's literate programming noweb) or IDE support like Elena described. For now my work flow is write the code backwards (ie: manually move the cursor up) and/or chopping and pasting. Then when I'm happy with it, I re-chop it all in my 'preferred' order and put a declare at the top. That sounds quite inefficient, but VI is really great for re- organizing text blocks so it is not too strenuous. That said, I'm really interested in ways that literate programming style can be followed with the least external support. Regards, Tim. On Mar 26, 4:15 pm, Mark Engelberg mark.engelb...@gmail.com wrote: On Wed, Mar 11, 2009 at 5:18 PM, Timothy Pratley timothyprat...@gmail.com wrote: It is also quite trivial to patch the compiler to auto-def symbols as it finds them instead of throwing an error. I would be interested in knowing how to do such a patch. When I work on code, I like to organize my functions in a way that makes it easy to read and understand what is going on. As I work on longer chunks of Clojure code, I'm finding that shuffling around the functions to avoid a lot of forward declarations is destroying the readability of my code. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Clojure group. To post to this group, send email to clojure@googlegroups.com To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en -~--~~~~--~~--~--~---
Re: Is Clojure's lack of support for forward referencing a feature or a bug?
:) I like that description! :) On Mar 26, 4:47 pm, David Nolen dnolen.li...@gmail.com wrote: For what it's worth I'm a big fan of the wishful thinking programming style. I write some functions how I think they should look, declare the functions I haven't defined yet. Then I implement the lower level functions, write some tests- then usually the higher level stuff works without too much tweaking. In a non-functional programming language this doesn't work so well, but it's a good fit for something like Clojure. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Clojure group. To post to this group, send email to clojure@googlegroups.com To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en -~--~~~~--~~--~--~---
Re: Is Clojure's lack of support for forward referencing a feature or a bug?
On Mar 11, 4:23 pm, quasar quasistellarli...@gmail.com wrote: It seems it makes Clojure source code to be in the order of lowest-to- highest abstraction. Naive mutual recursion based on top-level functions is impossible. I am curious, is it due to the current implementaiton of Reader or by design? You can use (declare something) --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Clojure group. To post to this group, send email to clojure@googlegroups.com To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en -~--~~~~--~~--~--~---
Re: Is Clojure's lack of support for forward referencing a feature or a bug?
Clojure does support forward referencing (if I understand your question): user= (declare a) user= (defn b [x] (a x)) user= (defn a [x] (b x)) user= (a 4) java.lang.StackOverflowError Note: (declare a) is a synonym for (def a) which works also. It is also quite trivial to patch the compiler to auto-def symbols as it finds them instead of throwing an error. That makes it hard to discover typoed symbols. One strategy to have the best of both worlds would be to allow auto-def until a non-def call is made, and at that point warn if there are any unbound. Not quite correct for a fully dynamic language. Regards, Tim. On Mar 12, 2:23 am, quasar quasistellarli...@gmail.com wrote: It seems it makes Clojure source code to be in the order of lowest-to- highest abstraction. Naive mutual recursion based on top-level functions is impossible. I am curious, is it due to the current implementaiton of Reader or by design? Best regards, Leonid --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Clojure group. To post to this group, send email to clojure@googlegroups.com To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en -~--~~~~--~~--~--~---
