New pedigree schema
Work in progress
Builds on top of !10 (merged)
Hard-wire some columns?
Do we define any columns at all in schema.json
, or everything in the pedigree? The latter seems most generic and flexible, but the former might give us some complexity.
If we define some columns in schema.json
, which ones exactly? Is there a clear boundery between hard-wired columns and custom columns? Seems arbitrary to me.
How to specify the order of columns?
In forms or anywhere. Should it be customizable as well? In the schema?
How to deal with missing data?
Are the following equivalent?
{
"name": "John Snow",
"gender": 0,
"consultand": false
}
{
"name": "John Snow",
}
In general I think we cannot require custom columns to be present, unless we force the schema extensions to have required: [...]
for all properties it defines.
Does resetting a date column (emptying it) remove the property from the JSON file or set it to the empty string?