The current data together implementation has a lot of prior terminology that needs to be reconsidered in the context of simply browsing a dataset. We should aim to reshape the metadata attribution features toward a simple browsing feature. We should aim for baseline feature that allows browsing of the datasets that have already been backed up. The goal should be to reduce the confusion prior to Tuesday's event.
A carry-over concept of "primers" is, for instance, a little confusing.
[x] rename to just "public record" where possible
[x] remove progress bars
[x] remove unattributed / attributed, just display a list of content
The current data together implementation has a lot of prior terminology that needs to be reconsidered in the context of simply browsing a dataset. We should aim to reshape the metadata attribution features toward a simple browsing feature. We should aim for baseline feature that allows browsing of the datasets that have already been backed up. The goal should be to reduce the confusion prior to Tuesday's event.
A carry-over concept of "primers" is, for instance, a little confusing.
/primers/
to/public-record/[id]