tagbody vs. defun
in [Lisp]
|
From: Slobodan Blazeski on 27 Oct 2009 16:20 On Oct 27, 9:03 pm, Slobodan Blazeski <slobodan.blaze...(a)gmail.com> wrote: > On Oct 27, 5:17 pm, Bigos <ruby.obj...(a)googlemail.com> wrote: > > > > > > > There's one thing that puzzles me most. What is best practice to group > > several functions together? For example in body of a loop. > > > (dotimes (x 3) > > (tagbody > > (format t "one ~a~%" x ) > > (format t "two ~a~%" x) > > (format t "three ~a~%" x) ) ) > > > Is tagbody the right way to do it or even for one loop I need to > > define a function. > > > (defun mytagbody (x) > > (format t "one ~a~%" x ) > > (format t "two ~a~%" x) > > (format t "three ~a~%" x)) > > (dotimes (x 3) > > (mytagbody x)) > > > I am puzzled because in Lisp code I have seen so far I didn't see > > tagbody being being used, although in other languages there are > > constructs to group statements together. > > As other posters stated you don't need a grouping statement > (dotimes (x 3) > (format t "one ~a~%" x) > (format t "two ~a~%" x) > (format t "three ~a~%" x)) > However every time you write blockish code like above there's a big > chance you're doing something odd. How about below: > (dotimes (x 3) > (mapc (lambda (s) (format t s x)) > '("one ~a~%" "two ~a~%" "three ~a~%"))) > Or even something like this: (mapcan (lambda (x) (mapcan (lambda (s) (format t s x)) '("one ~a~%" "two ~a~%" "three ~a~%"))) (enum 3 1)) cheers Bobi (*) (defun enum (count &optional (start 0) (step 1)) (loop repeat count for i from start by step collect i))
From: Slobodan Blazeski on 27 Oct 2009 16:36 On Oct 27, 9:11 pm, x...(a)unnamed.xach.com wrote: > Slobodan Blazeski <slobodan.blaze...(a)gmail.com> writes: > > However every time you write blockish code like above there's a big > > chance you're doing something odd. How about below: > > (dotimes (x 3) > > (mapc (lambda (s) (format t s x)) > > '("one ~a~%" "two ~a~%" "three ~a~%"))) > > That is pretty odd too. What about this? > > (dotimes (x 3) > (dotimes (y 3) > (format t "~R ~D~%" (1+ y) x))) Good but too .. hm simple. Maybe below: (mapcan (lambda (n) (apply (lambda (s x) (format t s x)) n)) (outer '("one ~a~%" "two ~a~%" "three ~a~%") (enum 3 1))) > > Zach (*) (defun outer (x y) (loop for i in x collect (loop for j in y collect (+ i j)))) (defun enum (count &optional (start 0) (step 1)) (loop repeat count for i from start by step collect i))
From: Slobodan Blazeski on 27 Oct 2009 16:53 On Oct 27, 9:36 pm, Slobodan Blazeski <slobodan.blaze...(a)gmail.com> wrote: > On Oct 27, 9:11 pm, x...(a)unnamed.xach.com wrote:> Slobodan Blazeski <slobodan.blaze...(a)gmail.com> writes: > > > However every time you write blockish code like above there's a big > > > chance you're doing something odd. How about below: > > > (dotimes (x 3) > > > (mapc (lambda (s) (format t s x)) > > > '("one ~a~%" "two ~a~%" "three ~a~%"))) > > > That is pretty odd too. What about this? > > > (dotimes (x 3) > > (dotimes (y 3) > > (format t "~R ~D~%" (1+ y) x))) > > Good but too .. hm simple. Maybe below: > > (mapcan > (lambda (n) > (apply (lambda (s x) > (format t s x)) n)) > (outer '("one ~a~%" "two ~a~%" "three ~a~%") (enum 3 1))) > > > > > Zach > > (*) Sorry for posting buggy code but something really sucks when moving code between groups and the listener here's the working version. (mapcan (lambda (n) (apply (lambda (s x) (format t s x)) n)) (outer '("one ~a~%" "two ~a~%" "three ~a~%") (enum 3 1))) (defun outer (x y) (loop for i in x nconc (loop for j in y collect (list i j)))) (defun enum (count &optional (start 0) (step 1)) (loop repeat count for i from start by step collect i)) Just to make myself clear the Zach version with nested dotimes is probably the right choice mine was just a little exercise in obfuscation through loop less code. cheers Bobi
From: xach on 27 Oct 2009 17:25 Slobodan Blazeski <slobodan.blazeski(a)gmail.com> writes: > Just to make myself clear the Zach version with nested dotimes is > probably the right choice mine was just a little exercise in > obfuscation through loop less code. When possible, I prefer obfuscation through FORMAT. Zach
From: Pillsy on 27 Oct 2009 18:41
On Oct 27, 3:40 pm, Bigos <ruby.obj...(a)googlemail.com> wrote: > > I'm going to second Tamas' advice about reading a Common Lisp book at > > this point. Since you have experience programming in other languages, > > I think _Practical Common Lisp_, available at > > http://www.gigamonkeys.com/book/, is the best choice. [...] > > [1] PCL is excellent; its only shortcoming IMO is its lack of > > exercises. > I have read first 3 chapters. But the problem is, I can't learn while > staring at the book. I need to experiment, and feel how things > work.Perhaps, as you say, the book would be better with some > exercises. Another book to look at is David Touretzky's _Common Lisp: A Gentle Introduction to Symbolic Programming_[1]. I remember reading it in parallel with _PCL_, and found its exercises quite valuable. I think it lives up to its title in terms of being gentle; you'll probably just want to skim through the first few chapters. HTH, Pillsy [1] http://www.cs.cmu.edu/~dst/LispBook/ |