The issue is that by discouraging participation in the community, you end up with a repository of old questions with decent answers (that may or may not be 'decent' anymore) and fewer newer questions with even worse answers.
My experience is that using SO is basically trying to find a middle ground between an answer that is too old to be relevant, but not so new that the effect of the stagnation of the community hasn't made the answer less than helpful.
Another problem is that newer answers merely provide an alternative, that's neither better or worse. For example [2] is much newer than the existing answers, but not "better" than any of them. It merely provides an additional possible solution that's appropriate in some uses cases, but not in others.
The problem of older/outdated answers is real, but also much harder to solve than people think it is.
The problem with this is that answers change, for some languages the way you did X 5 years ago might still be the best way to do X, but for another language it's no longer the best way of doing it. And while S.O has done a good job of moderating and make sure they are that high quality source of info, they caused a second order effect in that no one wants to post there anymore. After a while you're left with a bunch of outdated answers and no one wants to post updated answers because "it's just going to get flagged".
The authors example of any new JS answers being marked as dups and old jquery answers being provided instead is exactly what I am talking about. I hardly ever use S.O for JS questions anymore because everything is from 5+ years ago with examples in Jquery or ES5 that you have to translate to modern JS. Yeah those are high quality answers, if we were all still using Jquery.
S.O could have gone about producing the "searchable corpus of answers" in a different way that would have preserved the quality but not dejected all of it's users.
I've seen plenty of "20xx: Now we do Y instead of X" answers posted years after the original answer.
> No, there is no such feature, so I made a custom hotkey: Ctrl + `
And then a few years later Microsoft implemented the feature and made the default hotkey Ctrl + `
---
I felt a bit of trepidation editing a top answer with >1k votes, but I did it: https://stackoverflow.com/posts/43012779/revisions
Probably they need a deprecate-answer feature instead.
I have answered many old questions and never once has this been flagged; usually it gets upvoted.
I have 10 "necromancer" badges: "Answer a question more than 60 days later with score of 5 or more" – some of these answered questions that were over ten years old. And this excludes the answers which haven't gathered 5 upvotes (yet).
Answering old questions is absolutely worthwhile in my experience.
What SO didn't expect was that their real purpose was as feedstock for LLMs that will consume them and surpass them entirely.
Pretty common trade off in large tech these days
They want to be the first result for "X in Y" and also have that be correct.
But if I ask how to write getter/setters in C# in 2009 and get an answer that represents the state of the art and accept it, that's it. That's pretty much locked in forever. Despite the fact that 14 years later, that answer should be very different.
But you can't "re-ask" the question. No one is going to "unaccept" the answer. And if anyone does edit the answer, it's not guaranteed to be correct anymore.
You get stuck with a corpus of knowledge that gets stale.
Musks claims are not consistent with his behavior.
Often time it is like taking a time-machine back ten years ago or more. For example a lot of JavaScript questions get answered with "use JQuery" even for things now built in, and no new questions can replace them nor is the timestamp on questions meaningful.
SO's policy will be the death of SO. Nobody has a reason to participate, it is a super toxic community, and it is out of date. Much like Twitter it is a matter of WHEN something better will replace it, not IF it will.
I used to ask “interesting” questions to build that quality knowledge base, usually in less popular topics. At one point I had almost half the Emacs Lisp questions.
I eventually ran into this problem where I explained that I tried nothing before asking. I briefly debated with the admins, then said if they felt that way to delete all my questions and I would delete my account.
They agreed to delete my questions then quickly realized I had close to a couple hundred good questions then undid the delete.
Needless to say, I no longer feed the site.
Hopefully, with more advanced AI’s we can scrape the web, organize the knowledge, and have our questions answered, skipping the searching.
Programmers could quickly become proficient in modern C++, Rust, Go, Haskell, etc by “Pair Programming” with an “AI” assistant.
I would actually love this to be a thing, since if potential posters would read the Minimum Complete Verifiable Example page^1 it would cut down on so much of the "I am already frustrated and then my 'why no work' question was closed" drama
I wouldn't even need advanced AI to get it done, as the existing Eliza's "why do you think that is?" model^2 would be pretty close
1: https://stackoverflow.com/help/minimal-reproducible-example
2: https://en.wikipedia.org/wiki/ELIZA#/media/File:ELIZA_conver...
What is sad is that several communities (not just SO by any means) discovered ways to rebrand this ancient vice into a modern virtue.
To be fair to SO, I believe they have made enormous progress in the last several years. They've been making pragmatic compromises away from their early idealism in favor of the messy business of being a healthy community.
I really enjoy the StackExchange sites these days (for the most part).
It would be nice if there was some place on the Internet that tried to strike a middle ground.
It's undeniable that the original intent of Stack Overflow is as you described; both Jeff and Joel have been very explicit about why they launched Stack Overflow and what its intended purpose was.
But Jeff left over 10 years ago and Joel was never very involved in the day-to-day operations. Since Jeff left things have been rather directionless; for a long time many Stack Overflow employees disagreed about lots of things and the net result was that ... nothing ever changed.
There's still many people who subscribe to this view of Stack Overflow, but also many who don't. In my own rant about Stack Overflow[1] I complained that these sort of useful questions being downvoted or closed.
In the end, this lack of direction results in a weird mish-mash site where different people are operating under different assumptions about what the "correct" behaviour ought to be, and it leaves everyone unhappy.
They tried. It didn’t work - https://meta.stackoverflow.com/questions/354217/sunsetting-d... - the site is still there at https://stackoverflow.com/documentation/
> Find the best answer to your technical question, help others answer theirs
And here's one of the first snippets on their about page:
> Stack Overflow helps people find the answers they need, when they need them. We're best known for our public Q&A platform that over 100 million people visit every month to ask questions, learn, and share technical knowledge.
Stack Overflow's website says nothing about this when you're looking at it briefly. They highlight the fact that it's primarily a Q&A site. The people that insist that SO is _not_ a Q&A site are needlessly defending the gatekeepers that have ruined the site.
Finally, it's funny that everyone always says that SO is supposed to produce high quality answers, because it stopped producing any high quality answers around 7-8 years ago. The new answers on the site suck. My usual routine is:
1. Google question
2. Click 5-6 different SO links
3. Get pissed off because the answers are all crap
4. Go to the documentation
5. Spend an hour to find the one small doc reference that actually gives me the correct answer.
I wish people would stop propagating this lie that gatekeepers are necessary for "high quality". It's a lie propagated by gatekeepers to defend themselves.
Note how you have the check "thanks, I will keep these tips in mind when asking" before being able to proceed. This has been shown for years and isn't a new thing.
Maybe some things should be clearer; as I mentioned in another comment[1] I think Stack Overflow lacks a clear vision, but people are certainly given more information beyond the tagline.
You can take any tagline to its extreme.
[1]: https://news.ycombinator.com/item?id=34887632
> I wish people would stop propagating this lie that gatekeepers are necessary for "high quality". It's a lie propagated by gatekeepers to defend themselves.
Don't brand people are "liars" because they have a different view of things.
To defend this as "different views" when their brand clearly suggest otherwise is, I believe, a bit disingenous. To the parent commenter's point, the community friction and aggressive closure of "seeming" but poorly identified duplicates are compounded by often poor searchability on the site itself. So the issue really is the gatekeeping that occurs by moderators who are operating against the vision the site itself communicate (Ed: regardless of how consistently or inconsistently the vision is current or historically communicated).
So what do you expect? A "tagparagraph" to describe all nuances? You're all fixating on this tagline to an unreasonable degree.
As I said, plenty of people have been VERY clear about all of this right from the start, over ten years ago, and you get a full page before you can ask a question which describes all of it. Again, the communication could be a bit better, but it's not "disingenuous"; that's just a nonsense accusation; almost everyone on the site actually contributing answers subscribes to this view to some degree (there are some nuance with this).
Not wanting to answer low-quality questions or being roped in to being an unpaid programming tutor answering the same beginner questions over and over again is not "gatekeeping".
And here's the thing: I've been on the site for 12 years and 4 months.
Of course, that means that I mostly use SO as read-only. I've found that it just isn't worth it to use it any other way. Since a lot of other users share my attitude, it seems like a terrible way for SO to operate and attempt to create engagement.
1. Lots of material in question. Very simple unexplained answer.
2. Sparse material in question. Want to be multi part blog in the answer.
Neither of those are received well.
The key is to have a question that is good, helpful, and reproducible by others too (so they can try it out or find it if they have similar problems) and enough, but not too much detail in the answer. The answer is the tricky part because people get caught with not explaining anything because they already understand the domain in the question or explaining every nuance of the solution even though it is unnecessary for the solution itself.
But, there's almost no value in posting questions and hoping for an answer back. If that's what you want, then the right place is wherever the folks hang out who work on that technology - IRC, Discord, email list, etc.
https://stackexchange.com/sites?view=list#questionsperday - look at "questions per day" and the "answered%" (which is down from its peek of over 7000 questions per day).
The fourth most visited site on the network gets under 100 questions per day. The site that I used to be active on (back when I was active) got about 50 per day (which was completely manageable for a dozen people to read every question every day, fix the ones that had problems and close the unanswerable ones) and is now down to under 5 per day.
But Stack Overflow is still getting thousands of questions per day. Each question, unless its really good to start with and shows that it can be answerable and isn't just someone trying to get someone else to do their work is only a tiny blip in the overall firehose of questions.
At the start, when there were fewer questions per day and more people were interested in answering the questions and moderating the content - that is when the site thrived. But as it got more popular and the quality of the questions has gone down, people don't care about answering questions as much and those who worked to try to keep what is there interesting have left.
If you look at https://meta.stackoverflow.com/questions/422885/2022-a-year-... and pull up the google doc for the "community moderation in time" ( https://docs.google.com/spreadsheets/d/1upnX9UX8ab8rde8DrGnO... ) you'll see that most of the moderation tasks that were done by the community have dropped year over year over year. That makes it harder for the good questions to get noticed and get answers since the poorly written ones remain on the site longer.
And yes, I do believe that other sources with people more familiar with that technology are better places to ask. If there's a discord for it - working with the devs there is likely helpful. If its hosted on GitHub and the team uses Discussions there - that's a much better and more focused Q&A format for that technology.
Those sites (and they exist now) typically suffer from having enough people providing correct answers (unless they're the type who enjoy answering what a Null Pointer Exception is again and again). You can try seeing if https://www.coderanch.com/forums fits what you are after.
The library based model requires fewer people and so scales better - but it still scales. The issue with the library based model is that the curation work needs to be relatively low friction - something that Stack Overflow has been failing at.
You are contributing to the common pool of knowledge. As long as the question and answer themselves are good.
Is it a strawman if it appears (among other places) in the 2008 launch blog post of one of the two cofounders[1]?
> Every question in Stack Overflow is like the Wikipedia article for some extremely narrow, specific programming question. How do I enlarge a fizzbar without overwriting the user’s snibbit? This question should only appear once in the site. Duplicates should be cleaned up quickly and redirected to the original question.
Or here’s the other cofounder’s 2018 retrospective[2]:
> Stack Overflow ultimately has much more in common with Wikipedia than a discussion forum. By this I mean questions and answers on Stack Overflow are not primarily judged by their usefulness to a specific individual, but by how many other programmers that question or answer can potentially help over time.
Reasonable people may disagree on how well SO accomplishes that goal, but I think it’s well established that it converged on having it as a goal fairly quickly and has had it in sight for basically all of its existence—unlike other initial positions which are known to have changed[3,4].
In no way is this meant as an endorsement of gatekeeping[5]; it’s just that it’s never been not about giving you the perfect answer to your question, for any value of you including me.
(SO calling itself a Q&A site is perhaps a little tautological, given it more or less defined the term. But Q&A as envisioned there is, in particular, not[6] a forum, which would be about answering or discussing a specific poster’s question.)
[1] https://www.joelonsoftware.com/2008/09/15/stack-overflow-lau...
[2] https://blog.codinghorror.com/what-does-stack-overflow-want-...
[3] https://blog.codinghorror.com/meta-is-murder/
[4] https://blog.codinghorror.com/listen-to-your-community-but-d...
[5] https://blog.codinghorror.com/stack-overflow-none-of-us-is-a...
[6] https://meta.stackexchange.com/questions/92107/are-stack-exc...
Yes. As you mentioned the blog post is from 2008, that's 15 years ago now. Jeff hasn't had any input on the direction of stack overflow for more than a decade. From Jeff's blog post that you linked:
> I have not worked at Stack Overflow in any capacity whatsoever since February 2012 and I've had zero day to day operational input since that date, more or less by choice.
So I don't see how his blog post from 2018 has any place in the company's current mission. He says in the blog post that you linked that he had plenty of opinions, but he has chosen to not have any direct influence in the company for over a decade.
And Joel stepped down as CEO in 2019. And Joel's vision has clearly changed. This is from a blog post in 2019, which probably reflects what the current vision of SO has become after seeing what SO has turned into:
> In many ways Stack Overflow’s specific rules for what is permitted and what is not are obstacles, but an even bigger problem is rudeness, snark, or condescension that newcomers often see.[0]
So, yes. People are building a strawman by trying to say the rude, snarky, meaningless cruel remarks are necessary to maintain "quality" answers.
[0]: https://www.joelonsoftware.com/2019/03/28/the-next-ceo-of-st...
I would contend that the rudeness and snark are a symptom of failed moderation tools.
On USENET, if there was a user who was bothersome to you - you could kill file them. Plonk and they're gone from what you can see.
On Reddit, you can block a user and you won't see their comments again. Similarly on Twitter you can mute or block a user. And again on GitHub. On HN you can click hide and collapse threads from people you don't want to see.
With all of these, you can make an annoying person not annoy you.
On Stack Overflow, you can't block questions from a user showing up when you look at a tag. You can't ignore answers that they have given on posts that otherwise invoke Duty Calls https://xkcd.com/386/ . You can't prevent a user from commenting on your posts. You can't filter questions to those who have had some criteria.
The combination of all of this means that on Stack Overflow, there are only three ways to handle a user.
You can flag their comments or posts and have a diamond mod step in (note that asking poor questions again and again isn't diamond mod worthy) and note that that sometimes can take days or weeks to get resolved.
You can take your ball and go home. Quit stack overflow and go cold turkey. There are more than a few who have decided that the hassle of dealing with users there isn't worth the headaches.
You can make it uncomfortable for them to stay. Snark, veiled (and unveiled) rudeness, and similar comments. Not that this is an excuse for rudeness - but rather an explanation.
One of the things to lower the temperature and reduce the amount of rudeness is to increase the ability for a user to see what they want and be able to avoid interactions with people who they don't want to interact with.
If the python tag is getting too many new programmer questions in September asking to make a circle with a for loop, the experienced programmer can either close questions (note that some people regard that as being rude) until they run out of close votes, down vote (again, some people regard that as being rude) until they run out of regular votes, and then either ask "have you read the documentation for for loop" or leave.
Rudeness is often an established user running out of moderation tools.
> Stackoverflow is sort of like the anti-experts-exchange (minus the nausea-inducing sleaze and quasi-legal search engine gaming) meets wikipedia meets programming reddit. It is by programmers, for programmers, with the ultimate intent of collectively increasing the sum total of good programming knowledge in the world. No matter what programming language you use, or what operating system you call home. Better programming is our goal.
Some formatting that isn't quite clear there. First sentence is bold. "good" is italicized in the original too.
https://stackoverflow.com/questions/1003841/how-do-i-move-th...
Note the first revision and its author https://stackoverflow.com/revisions/1003841/1
Then look at https://stackoverflow.com/posts/1003841/revisions (both pages, but the key part is on the most recent one) with the closed, locked, and deleted actions and who did them.
The (for lack of better terms) Atwoodian and Spoloskian philosophies of how the site should be run and that fracture continue to this day with (excessively simplified caricatures) some looking to delete everything that isn't good so that they could concentrate on what was left and others trying to fix everything even though it isn't good to start with and a "we'll get around to it" for most of the content.
That the fracture hasn't been resolved and Stack Overflow corporate still hasn't really come down and said what they want it to be (though they want more "engagement" so it feels like it will be more Spoloskian - but they don't want the Atwoodians to leave).
And sometimes it just makes stuff up out of thin air...
At least it is easy for you to check that the code it gave doesn't work.
Actually, ChatGPT is going to be really good for helping you with finding the tiny detail in a large body of documentation, as long as their docs were published prior to 2020
I notice that the "gatekeepers" of different language communities span from hostile to friendly, so to tar the entire site of Stack Overflow as being the same smacks of prejudice, rather than fact.
So what language are you doing work in? Anyone who doesn't preface sentences of anger with the language of their worfklow - especially when the questions, answers, feel and brainshare of SO communities are divided by language - is offering an opinion of pure sentiment, rather than rational assessment.
I personally have no trouble believing that it isn't functioning as intended, but I haven't participated in asking or answering for a decade-ish so I don't really know.
If you are a reasonably senior developer there are close to zero questions that you have that should be asked on stack overflow. You already have enough ability to figure out most problems on your own. Or if you can't, it's most likely too niche for SO and you need to hunt out a friendly subject matter expert instead.
Alternatively, if you're a beginner, it's likely all the questions you have have already been asked, and you are SOs target audience. But not for asking questions, just for looking up the answers. Use SO like a differently organised Wikipedia instead.
Blocking SO is a really odd response, it's one of the best and most comprehensive well organised data sets of beginner to mid level programming information. Don't contribute, fine, but why stop using what's there. Similar to how only 0.1% of people contribute to Wikipedia, doesn't mean it's not still a valuable resource for the other 99.9%
This explains why, when I see SO in my search results, I think "that's not a place to get my question answered", then act accordingly.
If SO was an answer search engine, SO would be collating answers from answers.com, yahoo answers, reddit, etc into one platform. Clearly it doesn't, so your thesis fails.
The result isn't a higher-quality StackOverflow or Wikipedia, it's a narrower StackOverflow or Wikipedia. And with the subject matter StackOverflow covers, narrower actually means lower quality, because answers fall out of date quickly, and new problems arise quickly.
My most recent question on StackOverflow was about Valgrind on MacOS, and it was deleted quickly. I was pointed to a bunch of questions, which I had already looked at, which had only answers from before Apple made some changes that broke Valgrind. Last I checked, if you're running an up-to-date version of MacOS, everything StackOverflow has to say about Valgrind is useless. Quality? I think not.
With that context in mind it’s easy to extract value from it now, and either by allowing ChatGPT to adopt a process, or by building systems around it that enable greater agency (as they seem to be doing with Bing) these issues will eventually be overcome.
Great! And the answer to that question being "we didn't allow anyone to answer it, good bye", it's a pretty awful goal.
I am curious to know what they think about it
Fixed to: "The primary purpose of S.O. is to produce a searchable corpus of answers with few duplicates and pretty OUTDATED answers"
The primary purpose of S.O. is to produce a searchable corpus of answers with few duplicates and pretty high-quality answers. They're optimizing for being the first google result for "how to do X in Y language." They're not optimizing for being the best place for an individual to ask "hot to do X in Y language."
When you consider S.O. through that lens, I think a lot of how the site works (aggressive close-as-duplicate-ing, for example) makes more sense.