Names of sftrack variables (the big 4) about sftrack

[birderboone]

My personal thoughts. I think having a prefix the way sf does might be the easiest thing to circumvent the column name issues, something like sft_ or track_, that way we have little chance that column already existed in the users data.frame.

[basille]

My take on it:

• I'm in favor of a prefix, but that also opens more questions. I like sft_, unless we can get both track_ and traj_ separately for each type of object (after all, geometries and timestamps are different for locations and steps). However, this will only apply to timestamp, grouping and error, right? Not the geometry itself (I agree, better to leave it to sf itself), which might actually be confusing. Maybe we could also have a sft_geometry attribute, which simply retrieves (and sets) the sf_column one?
• I think we should leave the option to use whatever name the user wants (with arguments like timestamp_name, grouping_name, etc. in the function), as long as we provide sane defaults.
• If the column names are already taken, the function should fail, unless overwrite = TRUE, in which case it should proceed, overwrite existing columns, and still throw a warning.
• My preferences for default pairs of attribute names/column names are (or a variant with track_/traj_ as described above):
  "geometry": sft_geometry = sf_column
  "timestamp": sft_timestamp
  "grouping": sft_grouping
  "error": sft_error

[basille]

Closing this issue now that sftrack is developed and on CRAN.