How should we handle daylight savings? Although I've gone for UTC, I'm now questioning it as it's very difficult to account for daylight savings as not all countries observe them and those who do do not change on the same day, which itself varies form year to year! Perhaps the safest is to actually use TZ database names (e.g. Europe/London), which are location based. R can use these to convert dates to other timezones, accounting for daylight savings. It does mean a really long list of potential valid values in the schema though. Have a look at https://github.com/Infectious-Disease-Modeling-Hubs/schemas/pull/24 for a possible implementation of this proposal.
Are we making timezone required? I'm opting for yes for now.
Do we want any validation of valid characters in the model_output_dir property?
Couple of questions:
Europe/London), which are location based. R can use these to convert dates to other timezones, accounting for daylight savings. It does mean a really long list of potential valid values in the schema though. Have a look at https://github.com/Infectious-Disease-Modeling-Hubs/schemas/pull/24 for a possible implementation of this proposal.model_output_dirproperty?