search

brad-cannell / codebookr

Create Codebooks From Data Frames
https://brad-cannell.github.io/codebookr/
Other
25 stars 5 forks source link

Use label instead of description #12

Closed mbcann01 closed 2 years ago

mbcann01 commented 2 years ago

Currently, the documentation gives preference to the description attribute. I think this is mostly because the column attributes table in the codebook document says, "Column description". I think this is arguably a more informative choice of words than "Column label".

However, whenever a column includes a label attribute and a description attribute, the label attribute is given preference by the code. In other words, the text in the label attribute is displayed in the "Column description" portion of the column attributes table in the codebook document -- not the text in description.

Therefore, the way the documentation is written and the way the code works seem inconsistent to me. I think that leaves two choices:

  1. Change the documentation to give the label attribute preference throughout. In fact, we can probably just drop the description attribute completely.

    • The pros of doing that include consistency with other statistical software and possibly less confusion from users.
    • However, I still think "description" is more informative language than "label".

  2. Change cb_get_col_attributes.R to give preference to the description attribute over the label attribute when they both exist.

    • Pros: Keeps the more informative "description" language and the attribute name matches the name of the section in the column attributes table.
    • Cons: Might be more confusing for users.
mbcann01 commented 2 years ago

Just talked this over with Doug. We agreed to keep everything how it is except that description will win out over label if both exist. The idea is that if I have taken the time to manually type out a description, it should win out over whatever happened to be in label.