[esip-semantictech] ACDD vocab as RDF

[Mcgibbney, Lewis J (398M)]

Hi Jonathan,
Replies inline

On 11/6/18, 10:49 PM, "Jonathan.Yu at csiro.au" <Jonathan.Yu at csiro.au> wrote:

    This came out of work we've been doing to develop netCDF-LD. We are developing a standard and a toolkit to encode RDF graphs from netCDF (meta)data to start enabling data fusion, data discovery, and other use cases. However, we identified a few gaps.
    
    The challenge we hit was the many conventions used to encode information in netCDF don't have stable web identifiers (ie. URIs). These are often described in documentation but often not machine readable/check-able (though CF has a checker). Our WG identified in the context of netCDF, that CF and ACDD were important conventions to be able to describe netCDF-LD graphs with. There are others of course...
    
    We're at a point where we can generate these graphs, but there is limited infrastructure to support their identity (though for CF standard names - NERC and MMI have done some work in this area). 

> Is it fair to assume here that the prior work you are referring to is the NERC Vocab server (http://vocab.nerc.ac.uk/) and the MMI ORR (https://mmisw.org/ont/#/)? 
    
    Ultimately, we'd like to be able to provide SPARQL endpoints to accompany the netCDF files and provide semantic querying/inferencing. This could sit alongside their access via THREDDS/Hyrax for example. 

> Yes. I think that this is possibly where COR could be utilized. Just so you know, the ESSIP COR service is merely an ESIP-specific deployment of MMI's ORR. There is little or no development work which goes on for the COR. All of the work goes on at ORR and then we leverage it all when we deploy COR. 
With this being said, would you be looking to use COR as the platform for data persistence and access? If the case is yes, then I suppose we could look at what kind of tooling (and in which language) you need to facilitate this. Through the Swagger codegen capability, we can generate client SDK's for loads of languages so please let us know your requirements and we will see what we can do to facilitate.
    
    A new and parallel piece of work we're embarking on is to develop a tool that can take a set of netCDF files and output schema.org flavoured RDF graphs (in RDF or JSON-LD). 

> Is your intent for this code to be developed under ESIP (as you have been doing with your ACDD work)?
Thanks