Variables.csv vs. dictionary_table

[DougManuel]

1. Are we maintaining Variables.csv or migrating to dictionary_tables? If so, are we maintaining all the information
2. Do we need to maintain variableCategories.csv? Do we have a place for the information in this table if we are not keeping (now that WWMeasure info has been removed)?
3. What are the implications for sharing and validation schema?

[DougManuel]

1. We will migrate variables into three tables:
   • method: is performing an experiment or steps to make something (specimen, including a location, sample or measurement). A method is akin to baking a cake. What ingredients do you need? What is the amount of ingredients, and how do you mix and bake them? When baking a cake, you add 2 cups of flour to the dry ingredients, combine with the wet ingredients and bake at 200 degrees.
   • measure: is performing observation or describing a property of something (specimens including a location, sample, measure, etc.). Akin to describing a cake. How much does the cake weigh? How many calories and nutrients does it contain?
   • attribute: is a description of something. What is the name of the cake? What world region did the cake come from?

[DougManuel]

2. Yes. We will maintain variableCategories, but rename them categories (method, measure, or attribute).
3. Sharing and validation schema can be refactored for the table changes.

[mathew-thomson]

Currently in version 2, everything is in one big sheet called a 'parts list'. Are we thinking that a csv of the parts list (which includes some mapping of variables back to version 1) will replace dictionary_table and variables.csv? @DougManuel

[DougManuel]

Confirming @mathew-thomson note that in version 2parts.csv replaces variables.csv and variableCategories.csv. In parts.csv there is a field called partType.

• All rows in variables.csv become partType = measure, method, or attribute.
• All rows in variableCategories.csv become partType = category.