-
Notifications
You must be signed in to change notification settings - Fork 530
Description
As discussed in yesterday's Trusted Data meeting, we need to design an "Item Reviewed" (working title) metadata field and subfields. (This will be a "compound" field.)
We plan for the field to live in its own metadata block and should come with a name for that too.
I'll try to summarize the discussion below.
Related Dataset
We don't want to use the "Related Dataset" field for a few reasons:
- Sometimes the review will be of something other than a dataset.
- There is no structure to the field. It's just a text area.
Here's a screenshot:
Related Publication
We how the Related Publication field has more structure. It looks like this:
Here's the Relation Type dropdown:
One thing we'd like to avoid, if possible, is the poor UX of needing to copy the same information about the DOI into various fields.
Related Datasets (proposed)
We took another look at prototype by @vera of an improved "Related Datasets" field from this repo: https://github.com/vera/related-datasets-cvoc
Here's a screenshot of how it looks:
Unlike "Related Publication", users don't enter the same information over and over.
Questions
- What should we call the metadata block? Human readable and machine readable name.
- What should we call the field? "itemReviewed" comes from https://schema.org/CriticReview
- How best to link to internal datasets and external URLs that may not be datasets and may not even have a PID?
- When a user reviews an internal dataset, can we avoid data entry that we can look up, such as the title and author of the dataset? Can we store the PID in the database once? Can we avoid having them enter multiple forms of the PID, as they do with Related Publication?
Metadata
Metadata
Assignees
Labels
Type
Projects
Status