You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently our tool discovery harvesting pipeline and the resulting tool store present one version of the software, i.e. the latest stable release (or in absence of any formal releases, just the latest git master/main version). This latest version always receives the URI https://tools.clariah.nl/$identifier. This means the metadata is subject to change at any harvesting round. Any previous versions are currently 'lost' to the tool store.
I'm now considering whether we should explicitly harvest and keep all old software metadata versions. This is also in line with what solutions such as eScience's Research Software Directory are doing and important from a reproducibility/citability perspective. Each version would get a more persistent identifier https://tools.clariah.nl/$identifier-$version and the full history of versions can then be viewed in the tool store (if explicitly requested, the default index would still only present the latest stable release).
We could either actively harvest all historic versions, or take the easier route and simply retain and archive previous harvests.
Ideas and comments welcome.
The text was updated successfully, but these errors were encountered:
This has been partially implemented already now, all tools receive a version-specific URI (and a generic one that points to the latest version). Keeping historic versions has not been implemented yet though.
Currently our tool discovery harvesting pipeline and the resulting tool store present one version of the software, i.e. the latest stable release (or in absence of any formal releases, just the latest git master/main version). This latest version always receives the URI
https://tools.clariah.nl/$identifier
. This means the metadata is subject to change at any harvesting round. Any previous versions are currently 'lost' to the tool store.I'm now considering whether we should explicitly harvest and keep all old software metadata versions. This is also in line with what solutions such as eScience's Research Software Directory are doing and important from a reproducibility/citability perspective. Each version would get a more persistent identifier
https://tools.clariah.nl/$identifier-$version
and the full history of versions can then be viewed in the tool store (if explicitly requested, the default index would still only present the latest stable release).We could either actively harvest all historic versions, or take the easier route and simply retain and archive previous harvests.
Ideas and comments welcome.
The text was updated successfully, but these errors were encountered: