Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The decline of Stack Overflow (2015)

hackernoon.com

Post reply on HN
Print view View original

231–240 of 304 posts

Re: The decline of Stack Overflow (2015)

#231

Post by Veedrac »

reitanqild wrote: post #203
Veedrac wrote: post #135

Earlier quoted context omitted.

The new questions queue ( http://stackoverflow.com/questions ) should convince anyone that help vampires are real. Even if not, most experienced answerers regularly see literal copy-pasted homework questions (often several duplicates from different students) and people circumventing question bans to write yet more horrible questions. Help vampires aren't just real, they're everywhere. If they didn't exist, it wouldn'…

Then why don't they just make it just a little bit harder to ask the first question? Today this is one of the default privileges, and IMO that is asking for work.

What does "a little bit harder" mean? Any reputation requirement is out of the window because it would alienate far too many people from ever contributing. A screening process wouldn't work well because SO already puts too much burden on moderators (which is largely why people end up making rushed moderation decisions).

Re: The decline of Stack Overflow (2015)

#232

Post by gregmac »

CM30 wrote: post #210
tlogan wrote: post #24

Earlier quoted context omitted.

The SO is in decline. And decline is caused by its own success. - The SO was great because answers for noob questions where hard to find. At this point of time (thanks to SO and their knowledge database) there are so many other resources to get these noob questions answered or they are already answered. And some of these other resources are updated to reflect the most recent technology changes while SO answers are st…

> - Now people start asking hard stuff - but engineers (I intentionally do not say developers) who know answers on these questions are not interested in answering. I suspect part of the reason they're not interested in answering is because a difficult question requires a lot more work to likely get less points as a result. Meanwhile, answering a ton of easy questions can be done quickly and gets more points over on S…

> I suspect part of the reason they're not interested in answering is because a difficult question requires a lot more work to likely get less points as a result.

I have to admit I have fallen into this myself on both sides.

The (now) rare times I browse the site to find questions I can answer, when I run across one like this, its usually by a low rep user, so I think "hmm, I'll spend fifteen or twenty minutes sharing this, get maybe one or two upvotes, and if I'm lucky the original person will accept the answer" and usually deem it not worthwhile. Partly this is because the correct answer is often "don't do it that way, you should do x then you can do z and completely avoid your problem instead" which understandably people don't like to hear.

Re: The decline of Stack Overflow (2015)

#233

Post by bboreham »

bsder wrote: post #172

The bigger problem I have with SO is that it's answers are static --they are stuck in amber for eternity. This is not fine for something like Rust which has changed significantly within the last year or two. Many of the highest voted answers are now WRONG , but there is no way to dislodge them.

That question already has an answer here: http://meta.stackoverflow.com/questions/326095/please-unpin-...

(Actually many answers. But no resolution.)

Re: The decline of Stack Overflow (2015)

#234

Post by gregmac »

user5994461 wrote: post #167
sklivvz1971 wrote: post #34

Disclaimer: I work in the Q&A team at Stack Overflow. What people don't understand is that SO serves literally all developers. No matter what choices we make there will be upset people. Do we reduce moderation? Help vampires take over. Do we keep moderation as is? People complain we are too strict (or not enough). Do we allow only English? We are elitist. Do we allow SO in other languages? We are fragmenting the comm…

> Disclaimer: I work in the Q&A team at Stack Overflow. In that case, I have a serious question for you: What's your plan for refreshing the site on the long term? Lemme explain. The "community" is actively closing new questions as "duplicate" if anything similar is already present. At the same time, the old answer (that all duplicates are redirecting to) is often locked down for various reasons. The blocking prevent…

I think the stance on duplicates needs to be relaxed, with some filters/tools to help prevent higher rep users from seeing them during normal use.

For example, don't show high rep users questions from low rep ones. Let other low rep users answer the question independently. It encourages participation.

Let the site link duplicates together, but in a symbiotic way (even show highly voted answers from other questions as an aside.. Especially showing this on the old questions).

Votes (or even higher weighting on recent votes) on the question can let the community choose which "duplicate" is best, and this can change over time.

Re: The decline of Stack Overflow (2015)

#235

Post by DanBC »

x1798DE wrote: post #198
DanBC wrote: post #191

Earlier quoted context omitted.

If you ask a question and it genuinely has been answered elsewhere a friendly note "I think your question is answered here [url], is there anything that isn't answered by that question?" is useful. But "duplicate of [url]" often links to a question that has enough differences to not be a dupe. It's hostile, and it's a sign that incentives are out of alignment.

But just telling someone the answer might be available somewhere else doesn't direct future traffic to the canonical question, which is the entire point of the "marked as duplicate" flag. If your question gets marked as duplicate and the question it's a duplicate of is a different question, it is not very hard to get it "re-opened". Maybe people don't know that this is true, but I think the main people having problem…

> it is not very hard to get it "re-opened".

I'd be interested to see some statistics. I strongly suspect that when most newbies are told to fuck off that they do, in fact, fuck off.

Re: The decline of Stack Overflow (2015)

#236

Post by jlarocco »

nine_k wrote: post #52
marcelluspye wrote: post #8

Earlier quoted context omitted.

>SO isn't dying any time soon, and the content is still good. While I'd say this is currently true, answers on SO tend to stick around for a long time, and the 'marked as duplicate'-heavy behavior means that even if the best practice to solving a given problem changes, the top SO google result won't. While this isn't a _huge_ problem right now, with the way the site is set up I only see it getting worse, to the point…

Highly upvoted answers are usually converted into "wiki entries", so that they can evolve, extend and restructure with time. It somehow alleviates the problem, as long as the both sides (asking and answering) care to actually update it.

I'm skeptical that happens very often.

The only wiki answers I've seen have been hugely popular with dozens of answers and hundreds of votes. Maybe it happens outside of that, but it's definitely not how it "usually" works, IME.

Most of the questions I see on SO have maybe 0-5 answers, and the average number of votes is something like 2. If you're lucky there's an accepted answer and/or an answer with a clear majority of votes. I've even seen cases where the accepted answer is wrong, while a more upvoted answer is correct.

Re: The decline of Stack Overflow (2015)

#237

Post by kozikow »

I recently replied to SO question with a "New feature have been added that supports your use case: link" that was genuinely solving a problem.

Reply got removed by a moderator saying that I need to describe solution and can't just provide the link. I didn't bother. Currently the only answer suggests sub-optimal solutions and I am not replying to SO questions again.

It felt like a land grab - it wouldn't help the questioner, but it would help SO, by bringing more data to their platform.

Re: The decline of Stack Overflow (2015)

#238

Post by shagie »

achikin wrote: post #220
reitanqild wrote: post #207

Earlier quoted context omitted.

You never found a super-useful question that mods had tried to kill?

Here it is[1]. I've answered it six months after the first time I came across it. It is so specific that some people considered it nonprogramming and tried to close it. Am I complaining? No. For me, the main purpose of SO is to share specific programmer's knowledge that is not covered by documentation and can only be obtained by experience. The only way to share the knowledge is to communicate with the other people,…

Looking at its history ( http://stackoverflow.com/posts/3326955/revisions ) one can see all of the closes and reopens. This question has not been closed. And it might have been a better question to ask on SuperUser (or now Unix and Linux) instead.

Just because it is something that can be done in the process of debugging doesn't make it a programming question (the question "how do you edit a file" is not a programming question - even if the file is named "hello.c").

Re: The decline of Stack Overflow (2015)

#239

Post by flashman »

nine_k wrote: post #52
marcelluspye wrote: post #8

Earlier quoted context omitted.

>SO isn't dying any time soon, and the content is still good. While I'd say this is currently true, answers on SO tend to stick around for a long time, and the 'marked as duplicate'-heavy behavior means that even if the best practice to solving a given problem changes, the top SO google result won't. While this isn't a _huge_ problem right now, with the way the site is set up I only see it getting worse, to the point…

Highly upvoted answers are usually converted into "wiki entries", so that they can evolve, extend and restructure with time. It somehow alleviates the problem, as long as the both sides (asking and answering) care to actually update it.

SO could be the internet's technical manual if good answers were properly organised and maintained - if the point of providing answers was to document 'how to do programming' and you could only close a duplicate by explaining how to documented answer applies. But the incentive is get in fast, answer and move on to the next question.

Re: The decline of Stack Overflow (2015)

#240

Post by bootload »

"77% of users only ask one question, 65% only answer one question, and only 8% of users answer more than 5 questions."

Old SO user, started during the beta (user:2092), never asked a question. I don't bother with SO because a) sub-standard login (oauth), b) the time to answer a question correctly with sufficient details means it eats into time I can do other things; c) SO mods/response are mostly arseholes.

I also hate how original authoritative documentation is drowned out on google with crappy code examples or empty questions, locked, down voted or ignored.

Post reply on HN