Re: Quality check of ProvRDF

On Mar 15, 2012, at 8:48 AM, Stian Soiland-Reyes wrote:

> Hi!
> 
> I've gone through the ProvRDF page on my way to Boston.
> 
> 
> http://www.w3.org/2011/prov/wiki/ProvRDF
> 
> See changes: http://www.w3.org/2011/prov/wiki/index.php?title=ProvRDF&diff=6581&oldid=6536


FWIW, changes to the PROV-N signatures in the ProvRDF page such as:

|asnExpression([id],e,[a],[t],[attrs])
+
|asnExpression(id,e,a,t,[attr_1=val_1, ...])

and

|entity(id, [ attr1=val1, ...])
+
entity(id, [ attr_1=val1, ...])


break the automated alignment check at 

http://aquarius.tw.rpi.edu/prov-wg/provrdf-owl-coverage



If you want these signatures to change, please raise an issue against DM.

Thanks,
Tim




> 
> 
> I've fixed the uneven spacing problem of the tables, it was caused by
> intermediate spaces in the wiki markup, ie. instead of
> 
>> | blahg
>> 
>> | bluh
> 
> use:
> 
>> | blahg
>> | bluh
> 
> 
> I'll need to go through it again with a check against the OWL as I had
> an old version in my checkout.
> 
> Most notable changes are on derivation:
> 
> http://www.w3.org/2011/prov/wiki/ProvRDF#Derivation
> 
> 
> The prov:Derivation now includes prov:generation and prov:usage so
> that it can fully represent the complete DM statement.
> 
> Note that this presentation is particularly verbose, so I'm proposing
> to add a new syntax to triples that are inferred from the others, so
> that you can distinguish triple that come directly from the
> translation (for instance :e1 prov:wasDerivedFrom :e1 or the type
> signature, :ag a prov:Agent) from the directly inferred ones (such as
> :ag a prov:Agent or :e2 prov:tracedTo :e1).
> 
> Triples that come from inference rules should not be included, for
> instance for wasQuotedFrom we should not include :e1
> prov:wasAttributedTo :ag1. This is not representing the DM statement,
> this is a conclusion that you can draw from the represented
> prov:Quotation and its prov:quoter and prov:quoted.
> 
> 
> We still need to sort some things out with the class hierarchy with
> respect to attribute/involved-id 'inheritance' in DM .. for instance
> will the attributes on a prov:Start also apply to a prov:Association?
> DM does not seem to say so.
> 
> 
> -- 
> Stian Soiland-Reyes, myGrid team
> School of Computer Science
> The University of Manchester
>