20
\$\begingroup\$

There are 'Unanswered Questions' that contain good code and do not get closed, because saying 'looks good' is really a comment, not an answer. Example: how-can-this-simple-handler-in-nodejs-be-improved *

There should be a mechanism to mark the question as answered.

\$\endgroup\$
  • 1
    \$\begingroup\$ This may be related. \$\endgroup\$ – Jamal Sep 13 '13 at 17:34
  • \$\begingroup\$ Also related: Should "Your code seems fine" answers be acceptable? \$\endgroup\$ – 200_success Jul 8 '14 at 5:04
  • \$\begingroup\$ @Jamal good link, I hadn't seen that. I've been trying to eliminate the 70+ JS Zombie population, and for 40% of them I don't know the libraries, so I just offer formatting improvement and decoupling advice, whether the OP asked for it or not. \$\endgroup\$ – Viziionary Jul 26 '14 at 20:25
20
\$\begingroup\$

Good code doesn't mean nothing can come out of a peer review.

Reviewing code doesn't always mean shred the OP's code to pieces, and propose better ways to do the same thing. It could just as well point out why every right decision was the right decision to make.

I don't have an example at hand, but if Eric Lippert were to post some piece of the C# compiler for peer review (slightly unlikely), I'm sure Jon Skeet would be able to review it and explain what makes this code be good code, and that would still be an awesome review post from which many people could learn many things.

I don't think unanswered questions should be automatically marked as answered.

\$\endgroup\$
  • \$\begingroup\$ Very Good Answer!!! +1 \$\endgroup\$ – Malachi Nov 9 '13 at 17:05
  • \$\begingroup\$ Right. You can nearly always offer a way for the code to be more extensible or scaleable, via OOP practices, decoupling, etc. I've decided to try and take the JavaScript zombie population (currently 74) to zero, and while I don't know the library syntax for 40% of them, I can always answer based on structure and almost always formatting, even if I have to be nit-picky just to get the review done and partially kill the zombie. Crawlers are much easier to kill. \$\endgroup\$ – Viziionary Jul 26 '14 at 20:20
12
\$\begingroup\$

The problem is, unanswered questions are not necessarily good code. Either nobody bothered to write an answer or nobody had something to say about it so far.

Upvoted, unanswered questions can stand on their own as upvoted, unanswered questions. They are good questions, but no one knew an answer to it. There's no problem with these and I don't think there needs to be a mechanism to handle them. In this case (CR) I think that these questions stand for "hey, that's interesting code, it's well written and I have no idea what you could improve", there's no problem there. Some day somebody might come along and see something.

Not upvoted (zero-score), unanswered questions on the other hand need to be investigated, because there's something wrong with them. No upvotes and no answers can only mean one of two things: Nobody looked at them or they're not good questions.

I don't see the need for an extra feature to handle these types of questions. The good ones stand on their own, the bad ones already get reviewed.

\$\endgroup\$
6
\$\begingroup\$

I think that sometimes, if there isn't an answer, just giving your opinion on what you would have done differently and why is enough of a review.

this leaves a very fine line on CR though, and where do we draw that line. once I start nit picking the code for formatting the way I like to see it I usually find something else in the code that would be more readable if written differently.

also, I think that even though a question doesn't have the readability tag listed on it, we should still review for readability, so suggest what would make the code more readable if it were 100 times more complex, and they used the same formatting.

things like that are good answers I think

\$\endgroup\$
  • 4
    \$\begingroup\$ I agree. CR answerers are free to review any aspect of the posted code, not just what the OP wants to hear. \$\endgroup\$ – Mathieu Guindon Nov 8 '13 at 18:31
5
\$\begingroup\$

Your example of https://codereview.stackexchange.com/q/6809 is an oddball. I would say that it's neither good code nor bad code. It's code that has been distilled down to a nearly trivial test case, so that it doesn't really do much of anything useful, yet still triggers a bug in node.js.

The problem is that the question is mainly about a bug in node.js itself, not in the posted JavaScript code. Therefore, nobody will be able to satisfactorily resolve the problem in the context of Code Review.

Therefore, I've nominated the question to be closed as off-topic.

\$\endgroup\$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .