not sure how useful ids are in return value
Closed this issue · 2 comments
amsan7 commented
Does the average user care about ids? They're not very descriptive in terms of which corpus, transcript, or which participant. They're good for avoiding duplicates, if any of these variables happen to be repeated, or for use cases as described in this issue: #5. But I get the feeling users might want to communicate their needs using corpora_name, speaker_name, transcript_filename, etc. We could provide options for including ids and searching by ids for more advanced users. Thoughts?
mcfrank commented
what about adding return_ids = FALSE
as a default?
also - we could decrease the importance of this feature by unique-ifying the child names on import, e.g. if there are two Adam's, return Adam (Brown)
...