SKOS, vocabularies of values and vocabularies of metadata terms

[rowan.brownlee]

Hi all,

ARDC provides a vocabulary service named Research Vocabularies Australia (RVA).  RVA hosts vocabularies of values.  The AODN Instrument Vocabulary is an example of a vocabulary of values, including values such as "active fishing gear" and "cameras".  Each value is a label for a SKOS concept, and each concept has a definition and an IRI.  This vocabulary of values is  used by AODN to provide standardised entry of instrument information in metadata.

To date, ARDC has not sought to host vocabularies of metadata terms within RVA, choosing instead to focus upon vocabularies of values.

This looks to me like an example of a vocabulary of metadata terms, expressed as a collection of SKOS concepts.  For example, I assume that this metadata term would be associated with instances of measurements.

As one who has assumed that SKOS begins and ends with the expression of defined values, I'm interested in this approach to expressing vocabularies of metadata as SKOS concepts.

• Does this approach reflect common and accepted practice?
• Are there boundaries beyond which this approach would not make sense?

Here is another example of a vocabulary of metadata terms, available as a csv file, and from here, as json files. Might this metadata vocabulary be a candidate for SKOS encoding?  If so, would you encode the constraints (such as "units" and "type") within the concept resource descriptions?  Or would you separate those out as validation rules to apply to instances of data that have been encoded using this vocabulary of metadata terms?

bye

rowan

[Nicholas Car]

Hi Rowan,

One model's values is another models metadata terms.

Where you say you assume that the Concept acoustic_area_backscattering_strength_in_sea_water would be associated with instances of measurements, I could just as easily have instances of observed property, of which acoustic_area_backscattering_strength_in_sea_water could be one, and associate instruments with that. This is a matter of perspective.

SKOS can't differentiate between categories and instances, it just has Concepts within Concept Schemes. Your assumption is based on a conceptual model beyond SKOS, something like SOSA.

I think RVA should allow vocabularies of anything and, if someone wants to make that an instance list or a categorisation vocab, then that's outside of the direct RVA offering.

The NVS vocab you indicated, P07, is part of a set of NVS vocabs, some of which participate in a SOSA-like model (called PUV) that arranged things into instruments, observed properties etc. Here is an NVS vocab Concept that also has extended properties within the PUV model: http://vocab.nerc.ac.uk/collection/A05/current/EV_OXY/?_profile=puv.

Cheers, Nick

------- Original Message -------

--
You received this message because you are subscribed to the Google Groups "AVSIG (Australian Vocabulary Special Interest Group)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to avsig+un...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/avsig/a8a8da7d-1998-4530-8910-ace023ba5739n%40googlegroups.com.

[Rob Atkinson]

I agree that arbitrary distinctions are unhelpful (c.f. OWL "punning") - it all models of something, with varying expressivity...

IMHO the core factor here is actually the architectural role such vocabularies play.  This implies centralisation is problematic and unsustainable - but we also need to lead from some significant "nodes" to set the standards and mechanisms for federation.

Why do we define a "standard" term for something? - to create interoperability.  Between what? 

There is a clear lack of coherent architectural models that identify different implementation patterns for the creating, sharing and usage of such vocabularies in a "system of systems" context.

If you wish to contribute towards a common solution rather than generating yet another disconnected node in the ecosystem, please weigh in to the activity proposed in the OGC Spatial Data on the Web Working Group - its early days as we are focusing first on SOSA updates and OAS ready schema "building blocks" - but its open and ready for contributions and will be progressed during the next 12 months.

https://github.com/w3c/sdw/blob/observables/proposals/ObservablePropertiesRegister.md

Note this is in a branch, and the main repository is likely to be refactored in separate submodules - however its in a form where you can submit issues or PR for advancing the discussion.

Note also that governance is out of scope - this activity is about the enabler for federation - one (or more) reusable models for Observable Properties.  A simple SKOS profile for classification of properties would be a sensible contribution, and could be aligned with an OWL model (i.e. "punning") 

_______________________________________________________   kind regards Rob Atkinson Lead Ontologist, Knowledge Architect at SURROUND Australia Pty Ltd Address   Level 9, Nishi Building, 2 Phillip Law Street; New Acton Canberra Australia 2601 Phone     +61 419 202 973 Email      rob.at...@surroundaustralia.comWebsite   https://www.surroundaustralia.com Enhancing Intelligence Within Organisations delivering evidence that connects decisions to outcomes Copyrights: SURROUND Australia Pty Ltd is the copyright owner of all  original content and attachments. All rights reserved.  Confidentiality Notice: The contents of this e-mail are confidential to the ordinary user  of the e-mail address to which it is addressed, and may also be  privileged. If you are not the addressee of this e-mail, you may  not copy, forward, disclose, or otherwise use it or any part of it  or its attachments in any form whatsoever. If you have received  this e-mail in error, please e-mail the sender by replying to this  message.

To view this discussion on the web, visit https://groups.google.com/d/msgid/avsig/CVn4T8HsUs7o3MswJKDn8w8nwnsox6iRg0a-XjITY8TQia9GloyTctxFE_nWwHT6OIo8SDgw8jiKEa8h9O1eqctggVaW4Cvc-0fwxwbBq8I%3D%40kurrawong.net.

[Les Kneebone]

Thanks Rowan for raising this. You've just helped me work out what I found a bit odd about the NERC vocabulary server. The example you gave from there (Climate and Forecast Standard Names) I would refer to as an 'ontology'. Yes, it seems to be common practice to refer to ontologies as vocabularies, or more clearly as 'property vocabularies' or 'metadata vocabularies'. I'd like to see an end to the practice, but won't hold my breath.

Agree with Nick that it's a matter of perspective - SKOS doesn't have clear rules about range (meaning rdfs:range). A skos:concept is after all an instance of owl:class - we may treat it like a class. 

But I think using concepts as classes steers the endeavour away from one of the key intentions of SKOS: simplicity, especially with respect to managing "thesauri, taxonomies, classification schemes and subject heading systems", to quote SKOS reference.

As for whether ARDC, or RVA should support editing, registration and publishing ontologies (meaning NOT skos vocabularies), I don't see why not. The RVA editor (PoolParty) already has some capability there. Currently, the RVA CMS is not really set up to render properly any non-skos organisation system (e.g. the browse visualisation can only read a skos vocab). 

Cheers all,

Les

[rowan.brownlee]

Hi Les, Rob and Nick,

Thanks for your perspectives on this.  In short

• these are valuable initiatives
• demonstration of community take-up provides incentives for registries to implement
• if there's a sliding scale of differences between a resource that is simple (e.g. flat list of terms), and a resource that is complex (e.g. ontology), how may a portal assist the user to make sense of what a resource is, to understand, to inform selection?

Background & related notes follow

In providing a vocabulary service that includes a publishing portal, one of the challenges for ARDC is to provide something that makes sense to the users.  A portal that helps to explain each resource, helps a user make sense of a resource, helps them arrive at an informed choice about whether or not to select and use a resource.

Prompts questions about what information to provide on a resource landing page, and how to do that in a way that is sustainable.   What are the features of interest about a resource, its component parts, its internal relationships, its relationship to other resources, and where it has been used?  

What does the user want to know?  How is the information sourced?  How may provision of this information be sustained?

Rob's initiative sounds valuable.  To reach agreement on an approach to expressing and  communicating information about a type of resource.  Nick's initiatives with VocPrez have included the potential to enable resource & resource description syndication across registries.  There continues to be work around fair digital objects, CODATA, fair information profiles and others.

All valuable initiatives. Demonstration of take-up by communities would provide an incentive  for registries to allocate resources to implement support, enabling syndication of resources and/or their descriptions.   To provide an agreed expression of what a resource is, its features & relationships that are of interest to a user.

Les's comments about puzzling over a resource, working out what seems odd, and the notion of something simple at the core of SKOS.  How may it be made easier for the user to understand a resource?  To make sense of what it is, in order to decide whether or not to use it?

bye

rowan