Identification questions have these features:
- Many casual users ask them, don't want to read some rule about how to asking such questions before asking them, don't want to get downvoted/closed/deleted, don't want to answer following questions on comments. Just want to ask it quickly and get an answer quickly if possible.
- In many case the memory is very faint and it could even indeed be a mix of different films, it is really hard to make it clear. Identification with the most vague description is both the need and the fun for answering.
- The ansewr need not to explain why it is the answer, just the result is enough. It is just the best guess that pops up in the anwerser's mind with the given vague description. It's very normal that different people get different pop ups - the randomness and casualness is the essense of such Q/As. It is not about right and wrong, it's only about helping others to get back the title if possible. The purpose of the traditional explainations seems only to be make the answer more "SE-compliant".
- These questions are low bar questions and is vulnerable to spams. It's hard to discriminate spams from the vague questions from real user with real need. Gaining the normal reputation for such question will be problematic.
The above features doesn't fit SE model well, the result will be like the users feel it is hard to ask question and the reviewers feel there are overwhelming flood of bad questions.
So these questions seems to fit meta sites better. But there seems to be no talk about utilize the meta system.
- The best way is SE provide multi-meta feature. That is, for example, there can be
id.movies.stackexchange.com, both work under the site specific meta model (with some minor rule changes). Then many feasible SE sites can open their id meta sites.
- Without the multi-meta feature from SE, tag on meta is the interim solution. This will also contaminate the meta site. But meta users are more experienced user and they can use excluding tags to filter out the contamination. How to guide casual user to meta and make the experience not weird is a problem.
- No review/downvotes/close/deletion of such questions - go completely casual. Only let the SE platform to deal with the most serious cases.
- Reward for answerer: maybe a "identified movies" (means accepted answers) number beside user name, and also new badges.
- If there is too much historical data load to SE, maybe auto delete the ancient ones.
- Limited questions per user in a period, e.g. you can only ask one question per week or even per month or 12 questions per year- there is not that many movie to be identified for most people.
- No rules, just warn the user to provide as many details as possible - to get higher chance to get an answer more quickly - you get what you provide.
The essence is to exploit the crowd expertise of SE with a casual framework. And that casual experience is intertwined with existing SE culture seemlessly (hopefully - it may have some unseen serious problems).
The original post may not make the point and thinking clear:
- What decides the experience of SE to general public is the main sites, not the meta sites.
- SE have strict model and requirements for the main sites to keep a high content quality.
- Some questions such as identification doesn't fit well with main sites model. That is, to keep high quality, there are always rules which seems to be too much for some question genres, which make the user feel it too hard to ask and also induce too much reviewing workload. The negative impacts on both side seems to be unnecessary for some question genres. Letting low bar questions to gain reputations has negative impact like attract more spam posts. Go casual on such questions is a solution.
- It is not easy to lower the requirements for such question genres especially going casual on the main sites due to the general SE policy on the main sites. The result is a dillema whether to ban them. The result of the ban is that real need can't be helped any more and also decrease the value of the site to users.
- This is why it may fit meta model better - meta can be more flexible regarding rules and it will not affect the quality of the main sites. Also, there is no reputation gain from meta, that could reduce spam posts.
- The last problem is that it is not easy to let SE to provide the multi-meta feature, that's only the long term goal.
- Then the remaining problem for now is that, can active meta users bother the inconvenience of always have to use excluding tags to filter out those contaminations? Here is another potential advantage of meta. For the general public, a clear non confusing site is crucial, that's why using tag to put id questions on the main site could be problematic - they don't bother to use tags. But active meta users are experienced users, for them, flexibiliy/feature rich is more important than smooth and simplicity and using a tag may not be a serious problem. These id questions on meta can be migrated to an id-specific meta site in the future.
- Also note that meta does not have to be always only about the site. In theory you can just talk about the site on the main site, the reason for meta is to spin off distinct unpropriate question genres from the main site. You can also do that for any such genres - like identification - best if each genre has its own meta site.