[esip-semanticweb] posted initial version of ToolMatch data model

[Eric Rozell]

On Mar 1, 2012, at 6:11 PM, Lynnes, Christopher S. (GSFC-6102) wrote:

> 
> On Mar 1, 2012, at 5:29 PM, Eric Rozell wrote:
> 
>> The SPARQL would be really simple assuming we use some level of OWL 2 reasoning.  I think the use case should extend beyond "data accessibility" to include any attribute of a dataset.  There are probably some "explanation" tools out there that could handle all of these use cases.   In your example:
>> 
>> SPARQL query for what tools can draw a map of a dataset:
>> 
>> Assume you have a dataset URI (call it, myprefix:My_Dataset_URI)
> 
> Of course, we probably aren't starting from a dataset URI, but one of several possible identifiers, like a GCMD DIF Entry ID, for example, or a DOI.  SHould be able to add those as properties to the Dataset though, and add clauses to the SPARQL query to "join" them.

Sure, the URI is just the simplest case.  It's a good point that you bring up though.  Do we want to extend the use case for some simple level of dataset discovery?  Say we have a GUI that asks for a dataset ID, after you type in some ID, should the GUI ask "Did you mean the DOI or the GCMD ID?"  Obviously, there won't be much (if any) overlap in these identification schemes, but its theoretically possible with any two identification schemes.  We may even want a GUI where people can first search for datasets by name or textual description, and then after choosing a specific dataset, find tools that match.

>> 
>> PREFIX ...
>> SELECT    ?tool    WHERE    {
>>    myprefix:My_Dataset_URI    toolmatch:mappedBy    ?tool    .
>>    //alternatively:    ?tool    toolmatch:maps    myprefix:My_Dataset_URI   .
>> }
>> 
>> Explanation for matching tool:
>> 
>> Assume the SPARQL query finds the URI of a tool that matches (call it, yourprefix:Your_Tool_URI)
>> 
>> We can construct a service that explains individual triples, such as:
>> 
>> PREFIX ...
>> EXPLAIN    {
>>    myprefix:My_Dataset_URI    toolmatch:mappedBy    yourprefix:Your_Tool_URI    .
>> }
>> 
>> The result of this might be some provenance information that says (in provenance RDF form...):
>> "This tool (yourprefix:Your_Tool_URI) matches this dataset (myprefix:My_Dataset_URI) because the dataset has DAP accessibility and has a latitude variable and has a longitude variable."
>> 
>> The user would then have some starting point for knowing how to use the tool with the dataset.
>> 
>> Cheers,
>> -Eric
>> 
>> On Mar 1, 2012, at 3:09 PM, Lynnes, Christopher S. (GSFC-6102) wrote:
>> 
>>> Hey, Eric, thx for cleaning up the model!  When I get a free moment, I'll see if I can add some of our own instances to it.  Then I think it would be good to generate some sample SPARQL that satisfies the key requirements in our use case, i.e., for a given dataset, what tools can draw a map of it (and does it have to be accessed in a specific way for that to work)?
>>> 
>>> On Feb 29, 2012, at 1:03 AM, Eric Rozell wrote:
>>> 
>>>> Hi Chris (and others),
>>>> 
>>>> I took a pass at the ontology.  Namely, I cleaned up some domain and range restrictions.  I also added some cool features of OWL 2 reasoning that should make this more "Semantic Webby".  I've documented this here: 
>>>> http://wiki.esipfed.org/index.php/ToolMatch_Proposal_0.2
>>>> 
>>>> Hopefully the CMAP images help...
>>>> 
>>>> Moving forward on this, once we settle on a model, it would be great to start building a Web service where people can upload tools and datasets and identify new data attributes of relevance.  After upload transactions, the OWL 2 reasoner would reprocess the data, and data collection / tool compatibility can be queried from an additional Web service.
>>>> 
>>>> --Eric
>>>> 
>>>> On Feb 28, 2012, at 2:47 PM, Lynnes, Christopher S. (GSFC-6102) wrote:
>>>> 
>>>>> I know folks won't have time to examine before the telecon to look at it, but FWIW, here is my initial attempt in Turtle and RDF/XML.
>>>>> 
>>>>> http://wiki.esipfed.org/index.php/ToolMatch_Model
>>>>> 
>>>>> I'd like to use one of the upcoming Semantic Web telecons to go over the model if possible.
>>>>> --
>>>>> Dr. Christopher Lynnes     NASA/GSFC, Code 610.2    phone: 301-614-5185
>>>>> 
>>>>> 
>>>>> _______________________________________________
>>>>> esip-semanticweb mailing list
>>>>> esip-semanticweb at lists.esipfed.org
>>>>> http://www.lists.esipfed.org/mailman/listinfo/esip-semanticweb
>>>>> 
>>>> 
>>> 
>>> --
>>> Dr. Christopher Lynnes     NASA/GSFC, Code 610.2    phone: 301-614-5185
>>> 
>>> 
>>> 
>> 
> 
> --
> Dr. Christopher Lynnes     NASA/GSFC, Code 610.2    phone: 301-614-5185
> 
> 
>