CentreForDigitalHumanities / CentreForDigitalHumanities/Textcavator
Inform users about index versions
- Dominant language
- Python
- Stars
- 12
- Forks
- 3
- Avg merge
- 3d 3m
- Merged PRs (30d)
- 9
Description
Becomes more relevant if we implement #1078
We occasionally update indices to support new features or add more data. When researchers cite data from I-analyzer, future researchers may not be able to replicate their findings in a newer version of the corpus.
A minimal solution to this would be the following:
- When we inform users about how to cite documents (#726), let the model citation include the data of retrieval.
- Include some explanation in the manual that we occasionally update indices (and thus citation should include the date, earlier results cannot always be replicated. etc.).
For something more involved, we might consider some of the following:
- Allowing users to see a log of the changes to a corpus. This could be as simple as listing dates when the corpus was updated and possibly the type of change (e.g. updated with additional data, re-indexed from scratch, updated mappings only).
- Introduce semantic version numbers for corpora to make it clear whether changes are non-breaking (i.e. only add new documents or fields)
- Allowing users to switch to an earlier version of the corpus and explore it in the interface. Probably too ambitious. There are some ways to do this but it depends on the type of update.
- Allowing users to download the raw data from an earlier version.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.