Re: [ORG] Disposition of issues

Hi Joćo Paulo,

  On 20/02/13 18:54, Joćo Paulo Almeida wrote:
> Dear Dave,
>
> My comments to some of the issues follow inline. You and James have done a
> great job summarizing all this, and I think we're nearly there...

Thanks. Comments inline.

>> ## ISSUE-50 (break linkage with foaf:Organization and foaf:Agent)
>>
>> No change.
>>
>> As previously stated, and as reinforced by James, I see the linkage to
>> foaf as non-problematic. There is considerable usage of foaf, and use of
>> foaf:Organization has been non-problematic in practice. I would strongly
>> prefer to retain the current relationship.
>
> I will not object to keeping this relationship, so I think ISSUE-50 can be
> closed.

Great.

> However, there are implications that should be clear to users of
> the ORG ontology:  foaf:gender (which formally is a property of
> org:Organization because it specializes foaf:Agent) should be avoided by
> users of the ORG ontology for org:Organization. Would it deserve a note?

I don't think that's necessary. foaf:Organization has been around a long 
time and we don't see people publishing genders of organizations. I 
don't see a strong need to specifically warn of cases like that, seems 
to much like us commenting on foaf.

>> ## ISSUE-51 (should org:Post be a sub class of org:Organization)
>>
>> No change.
>>
>> As previously stated, this is motivated by existing usage of ORG and is
>> not problematic for ORG applications which do not wish to exploit this.
>
>
>> I would like to close ISSUE-50 and ISSUE-51 if Joćo Paulo will accept
>> this outcome.
>>
>
> I still believe org:Post should not be a subclass of org:Organization.
>
> If one is interested in Posts that can be "held by multiple people" than
> one could simply exploit the fact that an instance of org:Organization is
> an instance of foaf:Agent and as such may hold posts, i.e., a post may be
> held by an organization, which would address what Bart has discussed in:
> http://lists.w3.org/Archives/Public/public-gld-wg/2012Nov/0017.html
>
> Are there scenarios which clearly require Post to be a specialization
> Organization?

As we've covered several times before, there are cases in UK government 
structure where a post in one organization is held by another genuine 
organization, not simply a collection of people.

However, since we really need to get closure off I'll drop this.

I've have removed the assertion that org:Post is a subclass of 
org:Organization and changed the third paragraph of the subsection on 
Posts under [1] to explain that applications can still have something 
that is both a Post and an Organization. Which means that e.g. Bart's 
usage remains completely valid.

With this change are you content to close this issue?

>> ## ISSUE-48 (domain/range of org:reportsTo)
>>
>> I have removed the redundant unionOf assertion, as suggested by Joćo
>> Paulo, and added a comment pointing out that since org:Post is a sub
>> class of foaf:Agent it remains possible to have org:Post as the subject
>> or object of an org:reportsTo.
>
> If org:Post is no longer a subclass of org:Organization, than ISSUE-48 can
> be closed as you propose. Otherwise, just rollback.

Given the proposed resolution of ISSUE-51 I have rolled this change back.

> Unrelated to these issues, but related to a previous discussion (that was
> not captured in the issue tracker and that was not finalized), I think
> that the document should specify whether the properties:
> org:member (range Membership, domain foaf:Agent)
> org:organization (range Membership, domain Organization)
> org:role (range Membership, domain Role)
> Are functional properties. In my opinion org:member and org:organization
> are functional.

Agreed, I have updated the ontology and specification to reflect this.

> I would also like to see the more complete diagram that we have drawn
> included (in a non-normative part of the document, see
> http://lists.w3.org/Archives/Public/public-gld-wg/2012Nov/0026.html),
> because I believe it would be useful for users to grasp the ontology. Once
> the issues have been addressed, we can produced an updated diagram. (This
> is an editorial issue only.)

As previously discussed I prefer the non-normative diagram to be compact 
and relatively readable. It need not be complete. It is to help someone 
get their heads round the important parts of the ontology, not to be a 
complete reference.

I also have reservations about the use of UML notation which represents 
literal-valued properties so differently from object properties (e.g. 
org:classification v. org:identifier) on your draft diagram.

However the choice of which to use is partly aesthetic and I'm therefore 
a biased judge.

If you can produce a diagram that reflects the changes we have just 
made, and you feel is a clearer alternative then we can look at it and, 
if necessary, put it to a working group vote.

Dave

[1] 
https://dvcs.w3.org/hg/gld/raw-file/default/org/index.html#reporting_structure

Received on Thursday, 21 February 2013 11:30:34 UTC