agroportal / project-management

Repository used to consolidate documentation about the AgroPortal project and track content related issues.
http://agroportal.lirmm.fr
7 stars 0 forks source link

Is it relevant to add a property to encode the moment where an ontology became deprecated ? #379

Open hibadjebabria169 opened 1 year ago

hibadjebabria169 commented 1 year ago

Question: Knowing that an ontology can be marked as deprecated, is there a way to automate the edition of this property to encode the moment where an ontology became deprecated. Similarly, can we edit this property when a new submission is uploaded.

jonquet commented 1 year ago

This is the role of the "valid" date metadata property. This date shall more or less encode the moment where an ontology become deprecated.

syphax-bouazzouni commented 1 year ago

In fact, there is no need to have a deprecated attribute if it can be infrared from another one like a date. At least in the UI, the user does not need to fill in two properties(i.e deprecated and the date), he should only fill in the date and we will automatically infer/fill the deprecated attribute

hibadjebabria169 commented 1 year ago

The property deprecated can be true even the ontology in the prodcution phase which means it's not retired and it's still valid yet

jonquet commented 1 year ago

Those are the defintions for the Validity date property:

Let's assume that "deprecated=true" is the beginning o the "invalidation process" and status=retired is the end. It means that :

jonquet commented 1 year ago

Complete rules in https://github.com/agroportal/project-management/issues/381 when discussion is over.

syphax-bouazzouni commented 1 year ago

Also a slight proposition, what do you think of adding deprecated to status possible values , instead of having a boolean attribute for just deprecated? (my goal is to get rid of the complexity here of interdependent properties as much as possible )

jonquet commented 1 year ago

Well one could rapidly say yes... but in fact no.

Although it would be ok to have a new status "deprecated" but this will be less precise than the situation where deprecated=true and status=somethingelse Still I could cope with this...

BUT deprecated is actually part of the MOD 1.4 we do implement: https://github.com/sifrproject/MOD-Ontology/blob/master/versions/mod-v1.4_profile.rdfs#L381

In other words, in MOD we decided to keep both properties because one of them is super standard (owl:deprecated) and because "omv:status" can be used to refine... but no one actually put on paper the dependencies btw properties!