In recent discussion on Meta Stack Overflow (MSO), Stack Overflow Docs has been the subject of many questions. Is it a failure? Can it be fixed? Should we get rid of it altogether?
Here are some of the recent threads on it:
- Tearing Down the Structure of Documentation
- Is Documentation a failed experiment?
- Redesigning Documentation
In my time reading through answers to many of these posts, I came across this answer by Shog9, where he references a very interesting answer to a question about what Stack Overflow Docs is. Here is what it said:
It’s [Stack Overflow Docs] simply an attempt to fix the failed mission of Stack Overflow.
This site was designed to be the authoritative source of knowledge, but turned to be an endless conveyor belt with “fix my code” questions.
So, the goal of Documentation project is […] to make it achieve its initial goal.
You cannot deny the fact that Stack Overflow participants are focused on fixing particular problems for the certain posters. Which makes answers way too localized and essentially un-reusable. I declare this as a person that searches for duplicates a lot.
While Docs are imagined to be focused on the generalized solutions. Will they succeed this time — is a question.
- Your Common Sense
I was surprised to see this referenced and collaborated by a Stack Overflow employee, so I decided to ask about it:
@Shog9 From the answer you linked to, it almost seems like you are saying that SO is a failed idea also [Failed along with Stack Overflow Docs]. Am I missing something?
Here is what Shog9 said about it:
Sorta, @Caleb. Stack Overflow has generated an immense volume of useful, practical solutions to programming problems, and thousands of new questions are answered every day — in that sense it is a success. That success is no small feat! At the same time… An awful lot of practical questions end up never being asked here, or remain unanswered; the depth of knowledge for small tags can be pitiful. There is a growing volume of noise that makes it hard to find useful information. There’s a perception (somewhat exaggerated) that crowdsourced debugging is primary purpose. These are failures.
The danger here is that we’ve come to fear failure so much that we resist even trying to fix the problems lest we make them worse or break something else. That we become so protective of the areas in which we’ve succeeded that we dare not risk them by learning from the areas where we’ve failed, or even admitting that they exist.
- Don’t use Stack Overflow to debug your code. That is what Google searches, Chrome Dev Tools, and debuggers are for.
- Be very thorough in researching the problem before posting on Stack Overflow. Duplicates are incredibly common.
- Remember to use flags to help reduce noise.
- Don’t just ask a question if you are stuck. Post questions about things that make you wonder.
- If you happen to have knowledge in a small tag, contribute.
- If you discover something that has not been covered, create a question with an answer.
- Create questions and answers that are more generic. You can use the specific problem you have as an example. This allows the content you post to be used in more situations and not just one specific case.
Also, be sure to read Jon Skeet’s (of course, who else) blog posts on writing questions: