I'm in the process of adding write features available with the 2.2 api to my app. I'm debating whether or not to add support for posting new questions. As the api docs state:
Stack Exchange sites take question quality seriously, and many checks are run on the actual websites that can require various user actions be taken. In the API, any "low quality" checks that are triggered cause the write request to fail. This includes situations where a CAPTCHA or guidance text would be displayed on the websites.
It could be a good idea to collect some best practices for apps to use when posting. Are there any quality checks that apps should consider implementing client-side (e.g. post length)?
This Meta Stack Overflow post might be helpful: Heuristics for detecting a bad answer?