I have encountered the following situation several times on cstheory.StackExchange.com, math.StackExchange.com or MathOverflow.net.
- A new user asks a question.
- Someone answers, or someone points out that something is unclear in the question.
- The questioner says “Thanks a lot” in a separate answer without accepting the answer, or makes a follow-up post as a new answer to the question without updating the question.
Why didn’t the questioner accept the answer or update the question? The reason turned out to be that after the questioner posted the question using a cookie-based unregistered account, he/she lost the cookie for some reason. I guess that he/she can use the recover the access to the account if he/she had provided an email address, but
- he/she might not have provided an email address, or
- he/she may not know that it is possible to recover the lost access to the account.
(Admittedly I do not know how unregistered accounts and the account recovery functionality work, so I might be completely wrong about its functionality.)
This made me think that it might be better to encourage people to register especially when asking a question. A line in FAQ or the Ask Question page will be fine. To be perfectly clear, this question is not about encouraging people to register when posting an answer, with which the problem is less serious in my opinion.
On the other hand, it must be a deliberate design that StackExchange allows posting questions and answers without registration. Therefore, I am not completely sure if it is good to discourage people from asking a question without registering.
This discussion only tries to address the problem for new users who are unfamiliar with the site but careful enough to read instructions. Trying to solve the problem for those who do not read instructions is out of scope of this discussion.