Why is QA still not stackoverflow?

Allow me to provide Saturday sofa talk on the topic “How we can equip ...” in the form of an article, not a question, because the article will be read 10 times as many readers, and it focuses on all readers of Habr, and not only on the readers of his questions.

The article also has one simple suggestion on how to fix things with QA. It sounds like this: let's open access to QA to everyone, including ReadOnly. Plus a couple of small notes to him. I think that after this, things will get better there. And below - the details about what happened there since the birth of the section, what is happening, and why I think that things are going bad there (not like on StackOverflow) .

Details

On September 1, 2010, the creators of the site made a new QA section, which was designed to replace the shortage of the site StackOverflow.com in Runet and become the leader of high-quality answers to complex questions. No, they did not make such statements, but without words from the side it is clear why the site and Runet need a new section. Those sections that lead to Meilru suffer in terms of technical answers with a lack of professionalism of different depths and wildness. Yandex quickly closed its "Answers". A number of companies in Russia are trying to do this professionally, but they are not “on the ear”. The situation is better in Google’s answers, Quora is doing something, but everywhere there are not enough literate people to answer every difficult question so that they can take a firm place in the community esteem rating, as it happens on StackOverflow.com .

After all, there is every reason for this - there is a powerful and stable barrier of professionalism in the habr authorization filter, which has proven itself according to the results - the answers in the comments and the articles show that there are few random people, and those that do, do not want to litter false or just incorrect statements. And in the answers to mail, for example, false statements abound - there are very different values ​​and their support through feedback.
')
Therefore, the calculation was correct - to create a question and answer section on the basis of the habarawers community, where people tested by the filter could ask and answer.

At first, everything was fine, the first people reached out to the section, the questions and answers were, indeed, professional and experienced. But they were somehow not enough. A year has passed, another one - on the site a stable opinion has formed that in QA there is “no one”. Admins have given a number of indulgences in the answers - estimates can be made not for the first 3 days, as in the articles, those who are “labeled” can write, apparently, without restrictions on the number of messages in the answers. Questions began to mix in a couple of common tapes. But still this is not StackOverflow. Even close.

The picture was clarified by the counters introduced to articles and question-answers around October 2012. They showed that people in QA are not a myth, but it is clear that a very limited subset of users visit the section, approximately 10% of article readers. Not only answers writers, but also readers of questions. As a result, what does it mean that the question was read by 300 or, at most, 1000 readers? This means that among the 300 there are 1-2, who really know the answer. This is what happens. Difficult questions - 1-2 answers; simpler questions - 5-10, and some questions have no answers. In particular, because the questions and answers, apparently, also have the effect of news, as on the articles on the site - they are read the first day, and 10% of people from the first day - in the following days. That is, there is no constant pressure for answering questions when the right person is found in a couple of weeks.

To compare the difference between a question and an article on the same topic, we can observe the counters of the question and the article (the author has nothing to do with this article - just a case in point just today.

QA	post
in 19 hours - 500 views	in 13 minutes - 340 views in 18 minutes - 600 views

What is the difference?

Unlike Stackoverflow, we have a closed system for asking and answering. If I want to ask at Stackoverflow, I don’t need to write an article, go through a UFO contest, get an invite and ask a question. I just need to register. Later my rating grows, both the questioner and the respondent (professionals, help with the details, how these things go there).

In QA, we have a filter all at once. Ask a question can only past circles of hell. Let the questions from this also turn out to be well-grounded and professional - not everyone is going to ask a stupid question in order to get minuses into karma. But here we are cutting off a huge chunk of unregistered readers - the rest of Russia and the Russian-speaking part of the world from those 3 thousand (regular authorized) readers who thought up to ask a question. So, instead of a base of questioners out of 100 million people, we have a base of up to 30-50 thousand people, if they have an account and non-negative karma.

Of course, out of 100 million, a much smaller part will ask the desired complex and thematic question than a part of the registered users of the site. But still, the barrier of asking the question is unreasonably narrowed. What we have from this - we see after the expiration of a 3-4 year section. we see the parochial not even a forum, but a modest passage through which it is not even known who is.

What is the barrier for a site like Stackoverflow or Yandex. Answers?

1) I have a technical question. I am looking for sites that answer such questions.
2) Found several sites, but the question remains. I realize that in this place I will get an answer.
3) No registration? I register. Somewhere you can even through social networks, and on Habré you need, as already mentioned, to write an article, go through a competition, get an invite.
4) After the invite, I can write a long-awaited question.

It turns out a very complicated procedure to ask a question.

What is the situation with the answers to the questions?

For questions to appear answers, it is necessary that people who are able to answer, saw the questions. When do readers of Habr's articles see questions? It turns out that almost never. If they came to the question section on an occasion - yes, they will. They will see that there is no one there, in particular. Next time, they will enter very soon.

See those who read the tape with questions. Approximately 10 articles - 15-20 questions in a row, in chronology, without targeting. And even those who read similar questions will see. The site has a targeting mechanism - the module "Related Questions". It is designed for those who are looking for answers to your question. I came to the topic with a question - I saw similar ones, and maybe I found the answer there, if not in the first topic.

But there is nothing for readers and potential respondents to find questions. Logically, we need a module in the “Relevant Questions” sidebar, which would find not only similar articles, but also questions on this topic on the subject of the article. Then the readers, having seen the question, could come in and answer in 3 months. Now most often the answers are written in the mode of reading questions as news - in the first 3 days.

Now - who will answer? Yes, there is the possibility of answers registered with any karma. But Reed-Only and the unregistered ones cannot, although if they read the relevant Habr article, they could have answered something. And by the way, by the way, raise your rating in case you get an invite. That is, a person who does not have the right to write articles could get such an invite for the ability to answer questions.

If these 2 problems were solved - the relevant questions and answers from Reed-Onley, then the threshold of entry for answers would not differ from the threshold of other sites - you just need to register. To ask a question - the threshold for entering a question should also be reduced. The audience of Questions and Answers will now increase at least 3 times. As a side effect, there will be more interest in the site among Internet users and a crowdsourcing mechanism will appear in the development of the answers section. The professionalism will decrease - but the feedback in the form of a set of ratings will be preserved, and you can create a setting for users - turn off readonly users' answers if they become too annoying. Mark user status at the time of the response icon.

Who will do this?

The site holders have a plan for other urgent work besides introducing new mechanisms and rescuing the section (from ridicule). It is not necessary to ask them to do this complex and creative business. They could outsource this task to some group of developers to study the problem, offer a plan and implement the functionality with all the formalities of the NDA. The sum of all the work, taking into account the work in a large combat project, can start from several hundred thousand rubles. The analysis began in this article (getting up from the sofa).

What does the public think about this and question-response sites?