4 ms·

Everything in this article is categorically wrong and antithetical to every principle of good programming ever articulated. The only problem here, as others hav

by darawk 11y ago

Everything in this article is categorically wrong and antithetical to every principle of good programming ever articulated. The only problem here, as others have already noted, is that NPM allows people to delete published packages. Small modules are not evidence of a problem, and they certainly aren't evidence of an inability to implement these things on the part of the people depending on them. Why would I implement left-pad myself when there is already a well-tested implementation that I can install? Building up an ecosystem of tiny abstractions, bit by bit, iteratively and evolutionarily, is how we get robust, well-designed complex systems. We don't get there by everyone reinventing the left-pad function to sate some misplaced appetite for self-reliance. The author seems to make some arbitrary distinction between things that are 'large enough' to be packaged and 'pure functions' which are 'too small' to be their own modules, and I just couldn't disagree more. Tiny, pure functions are ideal modules. They facilitate the greatest degree of re-use, most clearly articulate what they ought to be used for, and stateless things are, in general, more composable than stateful things. There is no better unit of re-use than a tiny, pure function.

smrtinsert 11y ago
I wish I could say you were joking. At best it comes off as ivory tower idealism, at worst the complete naivete of someone who doesnt code.
- darawk 11y ago
  I'm not sure if you're saying this of my comment or the OP, but I code professionally every day in this style and it's working out just fine. And the smaller my modules get the better it seems to work.
  - rantanplan 11y ago
    It's not working out fine. You just haven't been around long enough to understand it. JS code is the most disposable piece of any infrastructure. In all companies(mine included) that I know of, npm and the JS jenga tower of hell is the most brittle element that breaks every fucking day. It's the constant pain you can count on being around. The stack and the dependencies are a moving target. Like.. every minute. If you had coded in any other language other then JS you would know that. Come to me after 5 years and tell me if your "professional" JS code you're writing today is used by anyone and then we'll talk. I'll be waiting.
    - darawk 11y ago
      Perhaps you should be looking inwards, rather than blaming your tools, for the reason your npm 'jenga tower of hell' exists in the first place. I experience no such pain, and i've been doing this for many years. And i've coded plenty in other languages. None with anywhere near as good an experience as NPM if you know how to use it properly.
      - rantanplan 11y ago
        Tell me the languages and package managers you're talking about. Don't worry, we ain't going nowhere. Think hard and take as much time as you want.
ORioN63 11y ago
I don't know why you were downvoted, but I agree with you 100%. Pure functions are indeed a good target for modularity and deserve proper documention and proper testing, at the very least. For the readers of this comment: * How many functions did you not commented last time you wrote some code? * How many functions do you leave untested? * How many functions did you wrote more than once? * How many "trivial" functions did you wrote that actually took you 3 hours, because it's actually tricky, so you checked other implementations and tried to wrap your mind around it. Check haskell's hoogle to a small sample of this concept.
toupeira 11y ago
> Why would I implement left-pad myself when there is already a well-tested implementation that I can install? You call 4 basic assertions "well-tested"? https://github.com/azer/left-pad/blob/master/test.js https://github.com/azer/left-pad/blob/master/test.js
- ORioN63 11y ago
  What if it was well-tested? What if it was well-typed? What if it was awesomely documented? Would that be ok? I'm asking, because a lot of the discussion in this post comes a single simple function being a dependency.
- darawk 11y ago
  a) I think that's a fine number of tests for a module this simple, which is itself an argument in favor of small modules. And I think it's four more assertions than anyone implementing this inline in a project would have. b) The details of this particular project are orthogonal to the philosophy of small modules generally. Whether or not this module is well implemented or well tested has no real relation to whether or not it is a good idea to compose tiny modules.
  - toupeira 11y ago
    a) It's about half of what I'd write for my own, where I would make sure to test all edge cases (e.g. input string being longer than padding length) as well as Unicode characters. Writing expectations is cheap, fixing bugs later isn't. b) I agree, but you were the one offering "well-testedness" as an argument :)
- robmcm 11y ago
  You are assuming well tested is referring to unit tests, rather than the wealth of developers depending on and running it in their projects daily. Don't forget how many different targets there are for JavaScript, even if you had a test environment with limitless budget, it wouldn't compare to having a popular package on NPM.
l1ambda 11y ago
"Sometimes, the elegant implementation is just a function. Not a method. Not a class. Not a framework. Just a function." --John Carmack
- tluyben2 11y ago
  He wouldn't put that small function in a package (or even a separate file) though.

by darawk 11y ago

smrtinsert 11y ago

I wish I could say you were joking. At best it comes off as ivory tower idealism, at worst the complete naivete of someone who doesnt code.

darawk 11y ago
I'm not sure if you're saying this of my comment or the OP, but I code professionally every day in this style and it's working out just fine. And the smaller my modules get the better it seems to work.
- rantanplan 11y ago
  It's not working out fine. You just haven't been around long enough to understand it. JS code is the most disposable piece of any infrastructure. In all companies(mine included) that I know of, npm and the JS jenga tower of hell is the most brittle element that breaks every fucking day. It's the constant pain you can count on being around. The stack and the dependencies are a moving target. Like.. every minute. If you had coded in any other language other then JS you would know that. Come to me after 5 years and tell me if your "professional" JS code you're writing today is used by anyone and then we'll talk. I'll be waiting.
  - darawk 11y ago
    Perhaps you should be looking inwards, rather than blaming your tools, for the reason your npm 'jenga tower of hell' exists in the first place. I experience no such pain, and i've been doing this for many years. And i've coded plenty in other languages. None with anywhere near as good an experience as NPM if you know how to use it properly.
    - rantanplan 11y ago
      Tell me the languages and package managers you're talking about. Don't worry, we ain't going nowhere. Think hard and take as much time as you want.

ORioN63 11y ago

I don't know why you were downvoted, but I agree with you 100%. Pure functions are indeed a good target for modularity and deserve proper documention and proper testing, at the very least. For the readers of this comment: * How many functions did you not commented last time you wrote some code? * How many functions do you leave untested? * How many functions did you wrote more than once? * How many "trivial" functions did you wrote that actually took you 3 hours, because it's actually tricky, so you checked other implementations and tried to wrap your mind around it. Check haskell's hoogle to a small sample of this concept.

toupeira 11y ago

> Why would I implement left-pad myself when there is already a well-tested implementation that I can install? You call 4 basic assertions "well-tested"? https://github.com/azer/left-pad/blob/master/test.js https://github.com/azer/left-pad/blob/master/test.js

ORioN63 11y ago
What if it was well-tested? What if it was well-typed? What if it was awesomely documented? Would that be ok? I'm asking, because a lot of the discussion in this post comes a single simple function being a dependency.
darawk 11y ago
a) I think that's a fine number of tests for a module this simple, which is itself an argument in favor of small modules. And I think it's four more assertions than anyone implementing this inline in a project would have. b) The details of this particular project are orthogonal to the philosophy of small modules generally. Whether or not this module is well implemented or well tested has no real relation to whether or not it is a good idea to compose tiny modules.
- toupeira 11y ago
  a) It's about half of what I'd write for my own, where I would make sure to test all edge cases (e.g. input string being longer than padding length) as well as Unicode characters. Writing expectations is cheap, fixing bugs later isn't. b) I agree, but you were the one offering "well-testedness" as an argument :)
robmcm 11y ago
You are assuming well tested is referring to unit tests, rather than the wealth of developers depending on and running it in their projects daily. Don't forget how many different targets there are for JavaScript, even if you had a test environment with limitless budget, it wouldn't compare to having a popular package on NPM.

l1ambda 11y ago

"Sometimes, the elegant implementation is just a function. Not a method. Not a class. Not a framework. Just a function." --John Carmack

tluyben2 11y ago
He wouldn't put that small function in a package (or even a separate file) though.