Open calummackervoy opened 3 years ago
Footprints may also be useful for writing data to Pods, something which has also been coming up thanks to #3
In the (relatively) short-term we could use Shapes to validate the data being saved to a dataset, e.g. that a Task is submitted with the necessary taskImplements
property which is the responsibility of the TaskActor developer
There are two main languages for shapes on the semantic web, SHACL and ShEx
It's like SQL constraints for data, but at the application layer rather than the database layer
Why not ontologies?
Good (long) reference on Shapes can be found here: https://ruben.verborgh.org/blog/2019/06/17/shaping-linked-data-apps/
I'll try and summarise it a bit below
In my work on #3 I needed to do a little reasoning on RDF type. I was surprised to see that even reasoning the transitive property required some manual referencing to ontologies. I opened #14 which details an issue using custom ontologies and how to automatically traverse it in this situation
Ruben summarises the issue we could face with these kinds of overlapping ontologies:
Using just ontologies, we'd need to have a custom reasoning to bridge the gap between them, which actually for common vocabularies like VCARD and FOAF we have been
Shapes may become the generic solution which we need to get around ontology inconsistencies
In the end, we don't want to be writing code with assumptions about what properties are in the data at all: https://ruben.verborgh.org/blog/2019/06/17/shaping-linked-data-apps/#shapes-apps