Re: ORG and reuse of vCard Ontology

Hi Renato,

On 08/05/13 03:50, Renato Iannella wrote:
> Dave, et al....
>
> As I am not a member of the GLD-WG, then I can only provide my views for this WG to make decisions...I would also hope technical/semantic architectural  issues take precedence over "deadlines".

Indeed, though there is always a balance to strike between the reality 
of available time and effort and desire to do the best job one can 
within those constraints.

> First, I was going to make comments by the LC Date, but due to our own semantic issues with the new vCard WD, we missed your dates.

Understood.

> Second, I was a little mislead (my mistake) with the vcard:VCard box in the diagram in Section 1.0, and VCARD appearing in the Normative References in the C.1.

Understandable. This was a late change. We originally had vcard:VCard as 
the range of org:siteAddress. This was raised as an ISSUE by other 
working group members due to vcard address modelling not being 
compatible with EU INSPIRE, which in turn affected the companion RegOrg 
vocabulary.

Our resolution of this ISSUE is make use of vcard a recommendation 
rather than a requirement and, as discussed, the phrasing leaves 
flexibility in how to do that.

> I assume the latter should be moved to C.2

Good catch, I have made this change.

> (and [FOAF] should be moved from C.2 to C.1 as it is actually Normative?).

Up until now we have taken the view that we are stating the relationship 
between foaf and ORG normatively but that ORG could in principle be used 
on its own. This may be an arguable point but my preference would be to 
leave it as it is.

> And perhaps the vcard:VCard box in the diagram should be empty.

It is greyed out. I was keen to reflect the "recommendation" to use 
vcard in the diagram.

If this is problematic I will remove it from the diagram. [A replacement 
diagram is in preparation in any case.]

> Third, and here comes my stronger view, reuse of ontologies is very rare, as it is "so easy" to create your own semantics. I would like to see W3C reuse more, and I note your Charter does the right thing:
>
> "The group will have to determine whether it is better to reuse existing widely-deployed terms such as foaf:name and dc:temporal, in their existing name space, or mint new URIs in a w3.org name space. Even if the group decides to mint new URIs, it should link them to equivalent concepts (using, for example, owl:equivalentProperty links) unless there are strong reasons not to."

Indeed. We've have gone out of our way to do so. Linking to foaf 
(despite  push back and arguments over that) and linking to PROV-O. We 
originally reused vCard as well and only moved that to an informative 
recommendation in response to the INSPIRE issue noted earlier.

> Will you be adding equivalent properties?

There are already relevant equivalent/sub property/class links for foaf, 
PROV-O and SKOS.

> Fourth (semantic comment), I note that foaf:Agent is specifically used for "persons" in your ontology - I assume you are aware that the semantics of foaf:Agent is much broader; "An agent (eg. person, group, software or physical artifact)".

We are indeed aware of that and deem it both appropriate and necessary, 
otherwise we would have used foaf:Person.

Specifically there are use cases, and existing data representing UK 
government structure, where a post in one organization is held by 
another organization (a committee), acting as a collective agent, rather 
than an individual.

We also have the notion of an OrganzationalCollaboration which is an 
Organization all of whose members are themselves Organizations.

> Fifth (and here comes the vCard plug to consider), the updated vCard now provides much more cleaner semantics when it comes to people, orgs, locations, and groups.

That is good news.

Given that vCard is still being finalized (I note the issue over 
namespace at least) I don't think we can publish a definitive mapping 
right now.

My recommendation is that as vCard is finalized we work together to 
define the mapping between vCard and ORG terms. Then in publishing vCard 
you include those mapping statements. Once that's done we can find some 
way to include the mapping statements back into the ORG ontology as well.

I don't know enough about W3C procedures to understand how that would be 
done. Whether we could simply add the mapping statements to the 
ontology, or would need someone to publish a Note defining the mapping 
or whether some update cycle on the ORG document would be needed. In any 
case there seems rethinking going on about how W3C handles vocabularies 
so hopefully this sort of thing will become easier in the future.

> So, for example, v:Individual would be a good candidate for a "person",

As noted about we actually need Agent not Person.

If you state that v:Individual is equivalent class to foaf:Person (which 
I assume is what you mean) then that would be sufficient to close the 
loop. That's not something that would be appropriate to state as part of 
ORG.

> and v:Organization = org:Organisation,

If that is what you intend for v:Organization then that's great and 
should definitely be added to ORG once the new vCard is stable.

I assume you are happy with the corollary that :

     v:Organization = foaf:Organization

> and v:Location for org:Site.

Is it?  I don't see in the current draft anything which defines the 
properties of a v:Location so it's hard to tell. If a v:Location indeed 
represents a site which could have an address, and if it allows for 
other address representations so that we don't undo resolution of the 
INSPIRE/RegOrg issue then that's fine.

Are you satisfied with this response?

Dave

Received on Wednesday, 8 May 2013 09:14:01 UTC