search

IGSN / jsonschema-testing

Playground for IGSN 2040 JSON Schemas
MIT License
0 stars 0 forks source link

Vocabulary for IGSN Core #2

Open jesserobertson opened 4 years ago

jesserobertson commented 4 years ago

As raised by @dr-shorthair in #1 (just seperating the semantics from the framing)

Todo:

jesserobertson commented 4 years ago

On existing concepts, I'll pull out an (abridged version) of a recent email trail so everyone can see what we're talking about:

My original email was asking about namespaces:

The other thing I was asking about was whether we could have a new area under schema.igsn.org for some JSON Schema documents for some stuff Doug and I are putting together. Currently the XML schemas reside at schemas.igsn.org/$version/foo.xsd - I’d suggest creating an endpoint like schemas.igsn.org/json/$version/foo.json for the JSON Schemas. We may also want to think about a basic vocabulary service (at least for our registration metadata) – maybe vocab.igsn.org? Alternatively we can publish JSON-LD definitions under schemas.igsn.org/json/vocab which might be easier for experimenting in the short term with an option of moving out to a separate vocab service at a later date if that works better.

Lesley pointed out that we have quite a few concepts in ANDS vocab services already.

We already have this IGSN vocab published on research vocabularies Australia: https://vocabs.ands.org.au/search/#!/?activeTab=vocabularies&pp=15&q=igsn

I noted there are a few issues:

Thanks for the pointer Lesley. There’s a couple of issues here - one is that this vocab isn’t in the IGSN namespace (and appears to be a mix of general and geoscience-specific terms) which may pose a governance issue (who owns/is allowed to update these? ANDS? IGSN? GA?). The other is we may want to be able to make rapid changes to schemas as part of this development and I’d hate to step on someone’s toes. Essentially I’d rather do a copy-paste of these definitions and then go back with a simple update for ANDS/GA to consider at the end. So I still think it’s worth publishing our own descriptions at least for the core registration metadata as a canonical source.

Lesley also pointed out:

The other issue that ARDC now has is that people are registering multiple vocabs on the same concept, so we are also now working through how we indicate to users which are the more authoritative ‘5-star’ vocabularies. (For example, until recently the only geoscience Vocabs in RVA were the IUGS ones – now Nick Car has now added the GSQ internal vocabularies…) This is pointing to something we need to consider with IGSN in that if IGSN does publish vocabs, we need to have a way of not just officially endorsing them, but also having a governance process that allows changes/additions to those vocabularies.

and Kerstin notes:

Which brings the question which vocabs IGSN should govern. We are leaving a lot of freedom to AAs on all other aspects of sample metadata. I guess we want to oversee all vocabs that pertain to the core metadata, but only sample_type is IGSNspecific. Material, collection method, locations, etc. will be governed by other organizations. I am not sure how this will work. For example, there are way more collection methods in SESAR right now than in any of the IUGS vocabs.

Not sure we're going to solve all of these but having something to start with would be useful.

dr-shorthair commented 4 years ago

I'm game to formulate the IGSN core as an extension to the SOSA ontology, if there is interest in taking that approach.

Where is the current state-of-the-art for IGSN core?

jklump commented 4 years ago

The current IGSN description metadata for environmental samples (a.k.a. "Passport") is at https://github.com/IGSN/metadata/tree/master/description. Included list enumerations are at https://github.com/IGSN/metadata/tree/master/description/include.