Hi Philip, I think Colin gave a good response further down the thread, in particular I think his statement "I at least have found Clojure has a much higher signal to noise ratio...<snip>...and can provide a significant reduction in incidental complexity, thus reducing the problem that some of these approaches are solving" is pertinent here.
I'll expand on my interpretation of that a bit. But first let me say that I'm pretty familiar with Uncle Bob, extreme programming and Agile development and the set of practices associated with those labels. I have been, in fact, a passionate advocate of extreme programming practices at times in the past. I do think there is a lot of value there, and I think every professional developer who has to work on teams and maintain code should seriously engage with that set of practices to try to understand the value. That said, there is also a lot of dogma and unquestioned convention in that community. I find Uncle Bob in particular to be frustrating in his rigidity, even when I agree with many specific points he makes. More relevant to the discussion at hand, since I've been using functional programming approaches, I've developed a more flexible attitude to what it means to "do development The Right Way." So first off, I would like to point out that I had no comments in my code--but I did document my functions with doc strings. I suppose the point about fragility stands, but on the other hand I am consistently frustrated when developers fail to document their public-facing libraries, and I think good documentation is a hallmark of high-quality software. So I guess what I'm saying is that I've accepted there is value in writing about your code vs. (or in addition to) simply letting your code speak for itself, despite its cost; the value outweighs this cost. The analogy image of the mislabeled container you attached, while clever, breaks down for me here: someone hasn't met all of their responsibilities. At the same time I do not disagree with the essence of what Uncle Bob says regarding comments: code that isn't clear and needs heavy commenting is *probably* problematic. However, I think the solution in functional languages is to lean heavily on the core persistent data structures and associated functions, use pure functions as much as possible and make it clear when you are not, specify types (in Clojure we can use schema or core.typed for this), and finally let the data do the talking. Regarding TDD: as I said I do use TDD in a professional setting, but not consistently: I don't think it is consistently necessary, and more than anything, the central argument of a lot of TDD apologists--that it helps guide design--seems much less compelling when using a functional approach. I haven't gotten enough clarity in my own mind as to why exactly this is yet, so I'll leave it at that--but let it suffice to say that I think TDD's biggest value in my day-to-day work in Clojure is in "automatically" providing a suite of regression tests, not providing design guidance. Finally, I would add that I agree with Kent Beck's points you listed at the end--I would simply argue that it's not necessarily the case that following Uncle Bob's stringent guidelines (for example) are the way to achieve those goals when using Clojure or doing functional programming in general. Thanks for making me think hard about this--I haven't tried to crystallize what I've learned since starting down this path a few years ago, and it's really valuable to think hard on it! Making high quality, maintainable software is challenging and deserves careful thought. Best, Dave (2014/12/09 9:24), Philip Schwarz wrote: > Hello David, > > I had set myself the constraint that I wanted the solution to exploit > two symmetries: > (1) The top left and top right of the diamond are mirror images > (2) The top half and bottom half of the diamond are also mirror images > >>I'm assuming you used a TDD process to write this (correct me if > wrong--basing that on the articles you linked to) > I was on a train commuting back home, and what I did was sit in a loop > where I wrote some code and then tweaked it until executing it in the > REPL gave me the part of the diamond that I wanted, by eyeballing the > console output. What a coincidence that in your gist you linked > to http://blog.jayfields.com/2014/01/repl-driven-development.html . I > was looking at exactly that blog post on Sunday to determine if what I > had been doing could be classified as REPL-based? Still not sure. Thoughts? > > My first version of the code was this > <https://gist.github.com/philipschwarz/c7e3be1ac97e482d04bf>: > > (defn print-diamond [letter] > (let [alphabet "ABCDEFGHIJKLMNOPQRSTUVWXYZ" > position-of (fn [letter] (inc (- (int letter) (int \A)))) > number-of-letters (position-of letter) > dashes (fn [n] (repeat n \-)) > fixed-text-for (fn [letter] (concat (dashes (dec (position-of letter))) > (list letter))) > template (map fixed-text-for (take number-of-letters alphabet)) > pad-with-trailing-dashes (fn [index line] (concat line (dashes (dec (- > number-of-letters index))))) > top-right-quadrant (map-indexed pad-with-trailing-dashes template) > top-left-quadrant (map reverse (map rest (take number-of-letters > top-right-quadrant))) > top-half (map concat top-left-quadrant top-right-quadrant) > diamond (concat top-half (drop 1 (reverse top-half)))] > (doseq [line (map #(apply str %) diamond)] > (println line)))) > > I showed it to Extreme Programming and Agile Guru Ron Jeffries, and the > following conversation ensued: > > > @philip_schwarz 1st stab at Clojure print-diamond using symmetries > identified by @TotherAlistair @RonJeffries @gdinwiddie @sebrose > https://gist.github.com/philipschwarz/c7e3be1ac97e482d04bf > @RonJeffries @philip_schwarz *can people read that and figure out what > it does? *i can't but not a closure person. @totheralistair @gdinwiddie > @sebrose > @philip_schwarz @RonJeffries @TotherAlistair @gdinwiddie @sebrose *I > like defns of top-half & diamond & think they r graspable-ish; > top-left-quadrant less so* > @philip_schwarz one interesting Q for us all is *if one didn't know the > prob could one grok the prog* @totheralistair @gdinwiddie @sebrose > @gdinwiddie .@RonJeffries I think *the program is generally easier to > grok if you've got the tests, too.* @philip_schwarz @TotherAlistair @sebrose > @philip_schwarz Dec 3 > @gdinwiddie @RonJeffries @TotherAlistair @sebrose agree - I have added > tests: > https://github.com/philipschwarz/diamond-problem-in-clojure/blob/master/test/diamond_problem_in_clojure/core_test.clj > > > I notice you did not write tests. I also notice that you added comments > to your methods. I like your comments. Find them useful. I am not saying > the following applies to your comments, but it will give you an idea of > the programming culture I am part of. In that culture, comments are > looked at with suspicion: > e.g. 1: https://twitter.com/nzkoz/status/538892801941848064 > > <https://pbs.twimg.com/media/B3qIJLFCcAEJLWm.jpg> > > > e.g. 2: "The proper use of comments is to compensate for our failure to > express ourself in code." - Robert C. Martin > e.g. 3: "Comments often are used as a deodorant... often ....comments > are there because the code is bad." - Martin Fowler > e.g. 4: > > * Primary Rule: Comments are for things that *cannot* be expressed in > code. > * Redundancy Rule: Comments which restate code must be deleted. > * Single Truth Rule: If the comment says what the code *could* say, > then the code must change to make the comment redundant. > > > In that culture, we aim to use certain implementation patterns that make > comments unnecessary. Also, where possible, the tests act as > (executable, more reliable) documentation. > > Moving on, after writing the terse first version of the code, I set out > to *make my code more readable*. > > Are you familiar with Robert Martin's dictum?: > > The Three Functions of a s/w module: > * The function it performs while executing > * To afford change. A module that is difficult to change is broken and > needs fixing, even though it works > * *To communicate to its readers. A module that does not communicate is > broken and needs fixing.* > > The rationale for making code more readable is an economic one. Here is > a brief summary of Ken't Beck's thoughts on the matter: > > o Economics is the underlying driver of software design > o Software should be designed to reduce its overall cost > o COST(total) = COST(develop) + COST(maintain) > o The cost of maintenance is much higher than the initial cost of > development > o Maintenance is expensive because understanding existing code is > time-consuming and error-prone > o Making changes is generally easy once you know what needs changing > o COST(maintain) = COST(understand) + COST(change)+ COST(test)+ > COST(deploy) > o Learning what the code does is the expensive part > o One strategy for reducing overall cost is to invest more in initial > development in hope of reducing or eliminating the need for change > o Such efforts have generally failed > o Beck’s strategy for reducing overall costs is to ask all programmers > to address the cost of understanding code during the maintenance phase, > by communicating programmer to programmer > o The immediate benefits of clear code are fewer defects, easier sharing > of code, and smoother development > > TO BE CONTINUED... > > Philip > > On Saturday, 6 December 2014 13:36:47 UTC, David Della Costa wrote: > > Hi Philip, > > I read your message and immediately wanted to try it myself--I > intended to leave it at that but I realized I would be remiss if I > did not give you a little bit of feedback based on my experience. I > should add that I was kind of fast and loose with my solution (that > is, I didn't really read the instructions), but it does print out > the diamond shape according to what I saw in the blog post examples. > > First of all, here's what I came up with: > > https://gist.github.com/ddellacosta/ba7e03951ba1bafd3ec9 > <https://gist.github.com/ddellacosta/ba7e03951ba1bafd3ec9> > > As you said, you weren't looking for alternative algorithms and I > recognize that that's not the point. But there are a few things > that I think are good and/or common Clojure practice that I think > I've internalized, and writing out an alternative solution helped me > to see them. > > - I'm assuming you used a TDD process to write this (correct me if > wrong--basing that on the articles you linked to), but I think a > repl-driven process may be more common for working through a problem > like this--i.e. something you can wrap your head around as a whole > and solve iteratively. That's not to say I and others don't use TDD > in Clojure dev, but just that it's also quite common to do a lot of > this kind of development in the repl. > > - you're grouping your side-effecting code w/the code that generates > the diamond data structure here: > https://gist.github.com/ddellacosta/ba7e03951ba1bafd3ec9 > <https://gist.github.com/ddellacosta/ba7e03951ba1bafd3ec9> > > While of course the diamond kata is a bit contrived and the point is > to print stuff out in the end, it also looks like you are trying to > be thoughtful about how you structure your code. So I would suggest > isolating your pure functions from your side-effecting code as a > sort of basic separation, and avoid monolithic functions like the > one I linked to above. This gives you the freedom to apply the data > structure to other processes if need be, rather than having to > refactor that code later on as soon as you need to do something > other than printing to the final diamond data structure. That is a > more compositional approach that is good to follow as part of > functional programming practice in general. And otherwise it seems > like you are following this approach--I think you can see this in > the shape of your code overall. > > - Stylistically, I found your naming conventions to be too verbose, > with not enough information about the actual input and output--I > would prefer a style like I used in my solution which aims for > readable conciseness, while documenting what is going in and coming > out of my functions. I assume Clojure developers reading my code > will have a good understanding of the core data structures and > functions available to manipulate them, and so I want to leverage > that as much as possible in how I write and document my code. > > In fact, at this point I prefer using Prismatic's schema > (https://github.com/Prismatic/schema > <https://github.com/Prismatic/schema>) to document as well as > provide further safety for my functions, and am of the opinion that > Clojure's one glaring weakness is its approach to typing--but that's > another discussion and I recognize this is not necessarily a > widely-held opinion. > > More generally, I think reasonable people could disagree on naming > conventions and so I would hesitate to say you're doing something > "wrong" here--I would rather say: the more Clojure code you read the > more you'll get a sense of how people tend to write. You'll figure > out what you want to adopt in your own style, and what Clojure devs > are going to expect. > > - I don't want to get too deep into the algorithm itself but I think > you would find it more natural to work line by line vs. the way you > constructed blocks and flipped them right/left, and you'd have less > code overall. I will boldly claim that my solution may be closer to > how other developers familiar with Clojure (or functional > programming in general) may approach it--not that I'm claiming it's > the best approach. I do think it is more concise without > sacrificing readability (which is subjective, I fully appreciate). > > - I don't know if I've ever once used a main function, and you don't > see them in libraries, certainly. But that is minor--there's no > reason *not* to use it, just that I wouldn't expect to see it. > > I hope this is useful feedback--good luck in your journey and enjoy > Clojure! > > Dave > > > 2014-12-06 19:48 GMT+09:00 Philip Schwarz > <philip.joh...@googlemail.com <javascript:>>: > > Hello, > > can you please review my first solution to the diamond kata [1] > and tear it to bits: let me know all the ways in which YOU would > improve the code. > > I am not so interested in a better algorithm for solving the > kata. I am learning Clojure and what I want to know is what YOU > would do to make the code more > readable/understandable/maintainable, or just to make it follow > Clojure idioms and/or conventions that YOU find effective, or to > follow a coding style that YOU find more effective. > > Thanks, > > Philip > > [1] https://github.com/philipschwarz/diamond-problem-in-clojure > <https://github.com/philipschwarz/diamond-problem-in-clojure> > > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to clo...@googlegroups.com > <javascript:> > Note that posts from new members are moderated - please be > patient with your first post. > To unsubscribe from this group, send email to > clojure+u...@googlegroups.com <javascript:> > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en > <http://groups.google.com/group/clojure?hl=en> > --- > You received this message because you are subscribed to the > Google Groups "Clojure" group. > To unsubscribe from this group and stop receiving emails from > it, send an email to clojure+u...@googlegroups.com <javascript:>. > For more options, visit https://groups.google.com/d/optout > <https://groups.google.com/d/optout>. > > > -- > 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 > Note that posts from new members are moderated - please be patient with > your first post. > 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 > --- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to clojure+unsubscr...@googlegroups.com > <mailto:clojure+unsubscr...@googlegroups.com>. > For more options, visit https://groups.google.com/d/optout. -- 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 Note that posts from new members are moderated - please be patient with your first post. 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 --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
