The question What hurts maintainability? recently got closed as being not constructive. In the comments two reasons were suggested:

  1. The fact that it has 18 answers is an indication that it's not constructive.
  2. wound up being a honey pot for creating a list of possible things that can hurt maintainability: lists aren't really what we do here.

What I'd like to know is why is it different from this?

It has even more answers, and pretty much invites the users to create lists, yet it also clearly gave people lots of value for their time, and the maintainability question seems to have done the same, judging by the votes. I personally don't see a good reason to close it.

share
2  
You'll notice in the history that the post you linked to was closed once and reopened by the community. I was one who voted for the reopen, mostly because it has a canonical answer, not because all the other answers are great. –  NickC Dec 7 '11 at 7:10
    
# of answers is a lousy litmus test. Not a bad hint that you should have a look and see if there's a good reason to close it, but not in and of itself a good reason to vote for it to be closed. The real problem, IMO is that the problem domain is frequently going to invite questions that don't have cut and dry answers. In fact, in this case, the more you boil an answer down to a list of general principles w rl examples, the better the answer is. I thought half the reason stack-programmers was created was to split the more generic advice-like questions away from stack-stack. –  Erik Reppen Feb 19 '13 at 17:02

1 Answer 1

up vote 7 down vote accepted

If you'll bear with me, there are two different answers: the official answer and the realistic answer.

The official answer

There is no difference between the two questions: the web developer question should be closed. We originally had a lock reason for this, but this was removed and we were told to close as not constructive any question that started to amass too many answers. The web developer question was closed, but was later reopened by the community.

Since it seems pointless to keep closing a question that can easily be reversed—even though that's the official policy—we haven't bothered to re-close it.

The realistic answer

The value in the Stack Exchange engine is that, for every question, there's going to be an answer that floats to the top. That answer has been vetted by the community (through voting) or by the asker (though accepting) as being the correct answer. So, when a visitor comes along in the future and has the same problem, they can just read the top answer and have their problem solved.

And so the major difference between the maintainability question and the web developer question is that in the web developer question, there is one answer that compiled all the knowledge from every other answer, effectively providing a definitive and comprehensive answer. You don't need to read any of the other answers because the first one covers them all.

That is, despite the question essentially being a poll, one answer is so damn good we overlook the question's other flaws. It's a rarity: an exception to the rule. One out of every hundred or so similar questions has the community actually take the time to meticulously maintain such an answer.

But 99 times out of 100, we just get a poll, where every answer just lists one small piece of the puzzle. If you wanted to even begin to understand the full answer, you have to read all the answers. But since none of the answers actually fully solve the problem, all the things we cherish on Stack Exchange—community vetting and voting—are meaningless.

Everybody's answer is equally valid and invalid at the same time because there's no way to judge what the correct answer would be: if you have an opinion about the subject, you technically provided an answer. Not really ideal at all, and not really what Stack Exchange is about.

That's what the maintainability question turned into. Despite deferring early flags and putting up a post notice to help keep the question on point, we still got junky answers we had to delete and nobody stepped up to provide a comprehensive answer. So when 3 members of the community voted to close and we got a flag pointing out what it turned into, we closed it.

Moving forward

Thankfully, there's a really easy workaround for getting questions like this to be on-topic and avoid the trappings of being labeled a poll and amassing a bunch of answers that don't quite cover everything: just provide a specific problem the answers can be tested against:

I'm working on project A, where I'm trying to do B. I want to make sure that when I pass it off to my client or the next developer, that all my bases are covered with respect to C. How can I make sure my project A is maintainable for the next guy?

Straightforward, to the point, and more importantly, testable. Someone can look at the situation, read the answers, and judge whether any specific answer actually helps the situation.

However, while this would've been helpful to prevent the state it's in now, what we have now is a ton of answers that don't really fully answer the question. We can fix that, too.

When a question automatically gets converted to community wiki (after getting 15 answers), it's supposed to be a not-so-subtle hint from the engine to the community that existing answers should be improved and/or consolidated: once a question goes community wiki, nobody gets any reputation anymore and post ownership gets heavily downplayed.

So while none of the 18 answers fully solve the problem, if we could create one comprehensive answer from all of them, then we'd be cooking with gas. If someone would like to do this, feel free to flag any answers that have been incorporated into the comprehensive answer and we'll delete them.

share
3  
I wish there were a way to explicitly encourage more of those awesome answers, even if they took time and were a community collaboration. I think that narrow questions look "easy" and encourage short answers that don't have much to argue with and therefore get upvoted by... everybody. –  NickC Dec 7 '11 at 7:19
    
@Renesis the auto-conversion to community wiki after 15 answers is supposed to be that encouragement, but most don't make that connection (I know I didn't until the whole too many answers = close thing came up). Once the reputation and post ownership incentives are stripped, it's a good opportunity to say, "okay, what have we gotten from these answers? Can we make one solidly awesome answer from what's here now?" –  user8 Dec 7 '11 at 7:24
2  
That was a fantastic answer, thanks. I'd like to add that I too find the use of community wiki to be very unclear, and I don't think that the post notice helped clarify what was expected of the users. We should consider to at least put a notice on all community wiki questions that say what the appropriate course of action should be. –  EpsilonVector Dec 7 '11 at 7:48
2  
"I'm working on project A, where I'm trying to do B. I want to make sure that when I pass it off to my client or the next developer, that all my bases are covered with respect to C." closed as too localized by (put your favourite mod here) ♦ –  user281377 Dec 7 '11 at 10:02
2  
Is the a ralistic venue to get this policy changed? Because it makes no sense whatsoever to me. Programmers.SE is for "conceptual questions about software development". Conceptual questions don't have a single objectively correct answer. And the workaround to "provide a specific problem the answers can be tested against" actually risks getting it closed as "too localized". –  Michael Borgwardt Dec 7 '11 at 13:46
2  
@MichaelBorgwardt The idea on Programmers is that questions could (or should) attract multiple complete answers that come from different perspectives. When each answer is self-contained and well-supported there is no issue with there being many of them and voting works as expected. We only run into trouble when a question attracts answers that individually are items on a longer list rather than self-contained. In that case there's no way to accept an answer and a future visitor would have to read through every answer to get help and that's what Stack Exchange is built to avoid. –  Anna Lear Dec 7 '11 at 14:34
1  
@MichaelBorgwardt It's fairly rare for a question to be closed as "too localized" in my experience. We might generalize it with an edit, but I don't believe the only possibilities here are "too broad to attract complete answers" and "too specific to remain open". There's a lot of middle ground there where a description of the situation the asker is in would form a great question. –  Anna Lear Dec 7 '11 at 14:35
    
@Anna Lear: I don't see a significant difference in usefulness or convenience to visitors between a question where voting determines a single, self-contained "best" answer, and a more open-ended one where answers complement each other and voting helps the visitor read the best ones. Stack Exchange seems to me to work just as well for the latter case, (except for the "accepted answer" feature which I've always seen as optional). –  Michael Borgwardt Dec 7 '11 at 15:11
    
@Anna Lear: And "multiple complete answers that come from different perspectives" sounds to me not very productive, since it implies either conflicting answers, with the format forcing the asker to decide on a "winner" even when the different perspectives may all have valid points, or redundant answers that all try to be the "most complete". –  Michael Borgwardt Dec 7 '11 at 15:14
1  
@MichaelBorgwardt Conflicting answers are okay, so long as both are backed up by facts and experiences. Then community voting should kick in to bring the best (overall) of those answers to the top. That's pretty much the essence of "good subjective". If some of the aspects of answers overlap, that's okay since each answer is still trying to make a separate point and give separate advice. If someone just has a minor addition to an existing answer, it should be posted as a comment instead of as a separate mostly-the-same answer. –  Anna Lear Dec 7 '11 at 17:50
    
@MichaelBorgwardt If you haven't before, check out these blog posts: Good Subjective, Bad Subjective and Real Questions Have Answers, and this Meta SO post linked from the latter post. They speak more about what subjectivity is about on SE and what makes a good (or bad) subjective question. –  Anna Lear Dec 7 '11 at 18:01
    
Has anyone put their hand up to consolidate all of the answers on this question into one answer? I don't have the rep on here to modify the existing accepted one.. I can just update my answer and consolidate everything. Any thoughts @AnnaLear? –  Deco Aug 1 '12 at 6:46

You must log in to answer this question.

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