Identify evaluation criteria needed in order to add packages to the collection
Just to add my own thoughts briefly, I think that this initial effort can cast a wide net, but we should settle on a slice of functionality for the purpose of a proof-of-concept. These initial criteria should give us confidence that our approach can accommodate the others, but they need not be implemented for a POC.
CRAN-like minimal quality standards, ideally fully automated (see R Consortium repo WG).
Initial submission requires peer-review with 'assessment vignette' including checks that can be rerun automatically and can be turned into human readable format ('documented evidence of. ...')
Continued monitoring based on metadata, e.g. sort out packages from new releases without reverse deps and low downloads, maintainer changes, ... Findigs can trigger re-review
Migrating cards from our meeting discussion
Just to add my own thoughts briefly, I think that this initial effort can cast a wide net, but we should settle on a slice of functionality for the purpose of a proof-of-concept. These initial criteria should give us confidence that our approach can accommodate the others, but they need not be implemented for a POC.