Spike: Feature Request/Idea: Deprecating metadata fields

[jggautier]

Overview of the Feature Request Dataverse installation administrators can deprecate metadata fields that are removed during a redesign of the deposit metadata model, such as re-designs of metadata blocks that Dataverse installations are using.

What kind of user is the feature intended for? (Example users roles: API User, Curator, Depositor, Guest, Superuser, Sysadmin) Superuser, Sysadmin

What inspired the request?

• Discussion about changes to the Citation metadata block to improve how Dataverse repositories collect funding metadata (https://github.com/IQSS/dataverse/issues/4859)
• Discussions that may lead to redesigns of other metadata blocks that Dataverse repositories are using, such as the "Geospatial" metadata block (https://github.com/IQSS/dataverse/issues/10398)
• Wondering if this discussion, about how to handle the removal of fields during a redesign of metadata models, can progress outside of any particular effort to redesign metadata models

What existing behavior do you want changed? In an issue about making changes to the Citation metadata block to improve how Dataverse repositories collect funding metadata, @scolapasta wrote:

We discussed this at tech hours and decided that the approach we would prefer (as it would generally be more flexible and it would maintain the concept of not changing old versions) would be to add an attribute to metadata fields so they can be marked as deprecated.

A deprecated dataset field would still allow old values to be viewed, but if you tried to edit you would be forced to no longer use that field (we would want some sort of instructions in the error message to advise you what you should use instead).

In this way, we won't have to write any scripts to change old values in the database. Rather we would break out this current issue into 3:

• one for the infrastructure component adding the new deprecated attribute and the logic for how to handle for create and edit
• a new API that would allow you to look up if a dataset's latest published version uses any deprecated fields (and possibly an API to get you a list of all such datasets)
• a script that could be used to change some of the old values we know to the new values we would like to use - these would be written in a way that they could be reusable for multiple datasets

Any open or closed issues related to this feature request? https://github.com/IQSS/dataverse/issues/4859

[cmbz]

2024/04/10

• Changed to spike issue due to large size. Requires analysis to scope work and create related technical issues of size 33 or smaller.