8

It is hard to see it not becoming what it is trying to fix (A repository of incorrect information). In 5 years, a lot of the top voted answers could evolve to be incorrect due to progressions in the programming languages, etc.

What I see happening is a ton of now incorrect answers being the highest-voted and accepted answers and being the first google result.

How will this issue be addressed?

3

3 Answers 3

3

For what it's worth, Joel attempted to address this issue when Stack Overflow launched.

You can also improve on the answers. If an answer is incomplete, expand on it. If an answer has a bug in it or is obsolete, you can edit it and fix it. Because Q&A in Stack Overflow are editable, you can safely link to a Stack Overflow permalink knowing it will always have a good answer. Stack Overflow won’t have the problem of other sites where obsolete or incorrect answers have high Google PageRank simply because they’ve been on the Internet for so long. If someone finds a security bug in an answer, it can be fixed… it won’t keep coming up in Google’s results for years and years poisoning future code.

(Emphasis mine.) In the interest of fairness, though, I should point out that some of what's in that post is no longer considered good practice on SE sites; for example, the suggestion of sticking on easy questions.

4
  • Yup, that is one way. However, when you have 'x' amount of years of questions that contain potentially wrong answers, the ability for the community to keep up with current Q&As AND the full history of Q&As will diminish. Jan 24, 2011 at 20:52
  • I would have a warning popup (or at least a different color on the date) for questions alder than, lets say, two years.
    – Trufa
    Jan 24, 2011 at 21:03
  • 1
    if you read through the chat we discussed some of the potential pitfalls to the solution in your post. I think this is an area where SE could be improved... Is there a chance that the solution posed by Joel (in 2008), could be outdated? Jan 24, 2011 at 21:05
  • @instance, I am aware of your discussion. I just thought the historical perspective would be interesting.
    – Pops
    Jan 24, 2011 at 22:05
2

Somebody gets really frustrated then and launches an own Q&A sites which will be populated with correct and up-to-date answers in a short time.

Then it degrades as well and the history repeats itself one more time.

It's happened with people, organizations and countries since the beginning of ages. Evolution, competition, rise then demise and a new birth. It will be the same with websites.

1

Programming is one of those fields which is nice about outdated information. You can take any C program on SO right now and compile it in Borland C 2.01 and it will likely run.

C, as a language, has not changed much over the years.

Some languages are still evolving, such as C# and their runtimes evolve as well, for instance the JVM supports a few important features now that it didn't a decade ago.

However, that version specific information is captured, largely, in version numbers, and those can often be found on individual questions, or if nothing else can be inferred by the date of the posting.

Further, in almost all cases these are backwards compatible, so the advice given now for a specific problem may not be the latest and greatest, but it will likely still work.

As much as the industry changes day to day, it's remarkably robust in terms of information age and relevance.

6
  • While I agree with some of your points, I disagree with the overall answer. Not all questions have version information and this isn't enforced. The majority of languages are still largely evolving, meaning that although it isn't as much an issue for C, it is for most others. It isn't just about language syntax either. A large range of questions could be considered 'too localized' between on how many years you apply to 'a specific moment in time'. stackoverflow.com/questions/8/… . Jan 24, 2011 at 22:09
  • Questions about concepts is a prime example of answers that can deprecate. Programming is still quite a young profession Jan 24, 2011 at 22:10
  • @Dan Most of the incompatibilities occur in frameworks and libraries, and it'll be blindingly obvious when one goes out to pasture, because another newer, shinier version will replace it. The language itself, however, simply doesn't change that much. C isn't the only language where this is true. The only languages that have changed drastically are those that are still in the early stages of research, and con be considered in, at most, a beta stage. Python, Ruby, etc are well beyond that. Many people are surprised when they learn how old Objective C is.
    – Pollyanna
    Jan 24, 2011 at 22:13
  • @Dan - Concepts will change over time as processing power becomes cheaper. But the concept remains the same - the only difference is that a new problem might have a better solution than it used to. The old solution is still perfectly capable of resolving the problem.
    – Pollyanna
    Jan 24, 2011 at 22:15
  • @Pollyanna, I think you miss the crux of the issue. Sure, things that don't work will be blindly obvious. What use is that though, when there are 14 blindly obviously answers to different questions that are wrong and 1 that is right (now). It is all about the signal to noise ratio, which is what SO was trying to fix. Jan 24, 2011 at 22:20
  • 1
    @Dan - I think you must help me understand. What is "wrong"? Do you mean, "Not the latest information, but it still works due to backwards compatibility" or "Will break program operation if used"? Further I think you're overestimating the problem. You say "a lot" of answers will be "wrong" but that is meaningless unless you frame the question. Are you estimating that over 70% of the content on SO will break people's programs in 5 years time? If so, I'd say that's a pretty urgent issue. If you instead mean that 500,000 questions will have out of date information, that's not a big deal.
    – Pollyanna
    Jan 24, 2011 at 22:24

You must log in to answer this question.

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