Should we get rid of cite string?

[kaitlynhair]

I know we have discussed this several times... but what do we think?
I've been thinking this it again as I've written a wrapper function for ASySD and the issue I'm coming in to is that cite_string isn't part of ASySD (label and source are already part of the functionality).

There are potential use cases for keeping it in certain scenarios e.g. comparing across strings AND sources at different SR stages. However, I personally think it does over-complicate things a bit e.g.the visualisations and tables. It will be a bit of work to remove it - but I would be happy to work on this if we agree to do so.

[TNRiley]

My original idea was just cite_source and cite_label. I agree that cite_string over complicates things. There are also no current outputs, plot nor table that use cite_string. The sankey plot on the plot sandbox vignette is the only thing that could potentially be integrated.

The only output is really something usable at this point would be an extra metadata element in the .ris/.bib/.csv export.

If we get rid of it now I think it could help keep things cleaned and lighter, but we could potentially lose out on expanded functionality.

I think the trade off is worth making ASySD easier to integrate. I also think we could work to build it back in if it's shown to be important to a future use case.

[TNRiley]

ALSO, just love to see the discussions area being utilized 🥇

[LukasWallrich]

Would it rather make sense to generalise ASySD, so that users can specify fields that just get merged in this way? That might make sense for URLs, database-specific IDs, retrieval dates, ...

I am not wedded to cite_string at all - but I can see cases where people might want to compare strings across sources and labels, and it would be good if the package (though not the Shiny) could accommodate that workflow.