For the TempAccount rollout, @TAndic has raised the issue of how [[ https://www.mediawiki.org/wiki/User_account_types | the new user type ]] will affect the [[ https://www.mediawiki.org/wiki/Extension:QuickSurveys#Defining_an_audience | "anons" audience configuration ]] for quicksurvey.
Open questions:
- find/confirm how the "anons" config is currently implemented (presumably using `User::isAnon()`?)
- is there an owner of this code (Trust & Safety Product, Readers Web) that is also part of the rollout?
- How does/should the new temp user account affect audience selection? Does Research have a preference?
- `User::isAnon()` == only reader
- `User::isTemp()` == user has edited
- `User::isNamed()` == permanent user (previously `! User::isAnon()`)
- Some approaches:
- leave "anons" config as is: only readers, where as previously users that ip edit were also considered anonymous
- switch to using `User::isNamed()` to keep previous semantics (temp editors are also considered "anon", which would make the config name incorrect)
- add new config param (e.g. user_type with values `reader`/`temp`/`named`).
# Requirements
[] This change is ideally done when no surveys are active. if this is not possible please coordinate with existing survey authors to make sure you don't cause disruption.
[] Solution should be minimal - ideally adding only one new field to audience to keep our code as simple as possible
[] Add validation for incompatible configuration e.g. if a survey is configured for names and anonymous users it should fail validation as this matches no users.