published -> discoverable, immutable

[alvacouch]

If something is published, it should behave as if the immutable (and discoverable) flags have also been set. Making something published should not be possible to undo except as an administrator. Undo should reproduce the exact flag state before it was published.

[alvacouch]

published -> immutable now. Not sure that it should imply discoverable. Question is whether public and published should be orthogonal?

[dtarb]

Lets avoid embedding limiting policies in the functionality of this library. Therefore I suggest that published be orthogonal to discoverable and public in terms of functionality. We have

• the most common use should be that once a resource is published, it should be discoverable and public
• there may be a case if we support or partner with a subscription journal that a resource that is published not be public, but that access only be limited to subscribers. This is not a use case that we want to actively pursue in the near term, but is not a capability we want to shut out which would be the case of public was implied by public.
• discoverable is similar to public. If a resource is discoverable it can be discovered by open searches. A subscription journal may want discoverability to only be through reference in the journal. Again this is not a use case we want to actively pursue or necessarily even encourage, but lets not rule it out by tying published to discoverable.

By leaving the setting of public and discoverable flags to the interface all the above seems possible. It might be possible to simplify and encourage a resource being made public once published by having the API call default to setting the public flag, requiring a specific argument to retain current settings of public and discoverable.

[alvacouch]

Currently public implies discoverable and published implies immutable. should I undo that?

[dtarb]

No. This is correct.

[horsburgh]

I think published should imply discoverable - and that should be part of the agreement when a user chooses to publish a resource. I can understand (but don't really advocate for) the use case of limiting access to published content (e.g., the journal subscription use case). But, I don't understand limiting access to discovering that a "published" resource exists. Why would we want to limit access to metadata? If we don't want to actively pursue this, then why don't we just implement it the way we want it to work?