W3C home > Mailing lists > Public > semantic-web@w3.org > May 2013

Re: New vCard Ontology draft

From: Adrian Pohl <uebertext@gmail.com>
Date: Fri, 3 May 2013 10:04:43 +0200
Message-ID: <CAOtd05GLLcgOS-cbe_oBhgNe8urpX7f198OxBfmhKgxbOkP8jQ@mail.gmail.com>
To: SW-forum Web <semantic-web@w3.org>
Hello,

nice to see that vcard is being worked on. Here are some random
comments that came to my mind when I took a look at the vocabulary.

- First, I am happy that the vcard:VCard class has been dropped and
the properties are directly attributed to individuals, groups or
organizations. That's following the pattern people have followed
anyway.
- Where can I download an rdf file of this draft?
- Object properties and datatype properties seem to be clearly
seperated by naming the first with a "has" in the URI ("hasName",
"hasEmail" etc.) and the second without a "has" ("gender", "email").
That's very helpful.
- I don't get it though, why there sometimes is the redundancy
providing a datatype and an object property for the same information
(email, geo, related). It seems to be intended that they are used
together like in the example

<http://example.com/me/corky>
    vcard:hasTelephone [ a vcard:Home,
               vcard:Voice;
               vcard:telephone "tel:61755555555" ] .

On the other hand, this bnode construction isn't necessary. The
example also says:

<http://example.com/me/corky> a vcard:Individual;
    vcard:email <mailto:corky@example.com> .

instead of:

<http://example.com/me/corky> a vcard:Individual;
    vcard:hasEmail [ a vcard:Home,
               vcard:Voice;
               vcard:email <mailto:corky@example.com> ] .

If these alternatives are provided deliberately (which might make
sense if you want to put personal and business information into one
vcard instead of having two seperate vcards), it would be nice if this
was made clear in the example section with use cases for choosing
either of these possibilities.

- Also: Shouldn't the range of vcard:email be xsd:anyURI if it is
intended to have the mailto: URI as range?
- I am with Masahide when he criticizes the re-naming of properties.
E.g. vcard:country is to replace the former vcard:country-name. That
makes no sense in two ways: Firstly, usages of the two vcard versions
aren't compatible and second - as the range of this property isn't a
URI for the country but the name string - it makes more sense to name
the property "country-name".
- A presentation or labeling issue: The naming of the properties and
classes in the HTML isn't always as the URI suggests. E.g. it says
"member" while the doc is talking about vcard:hasMember or "vCard
Kind" where it is about vcard:OrganizationName.

That's all for now.

- Adrian

On 3 May 2013 04:57, KANZAKI Masahide <mkanzaki@gmail.com> wrote:
> Existing RDF vCard (2010 member submission) is already deployed quite
> widely: the 9th most popular vocab in LOD cloud [1], and referenced by at
> least 6 othe vocabularies in LOV [2].
>
> Changing property names would break them. Terms should be stable, especially
> in the same namespace, IMHO.
>
> cheers,
>
> [1] http://lod-cloud.net/state/
> [2] http://lov.okfn.org/dataset/lov/details/vocabulary_vcard.html
>
> --
> @prefix : <http://www.kanzaki.com/ns/sig#> . <> :from [:name
> "KANZAKI Masahide"; :nick "masaka"; :email "mkanzaki@gmail.com"].
Received on Friday, 3 May 2013 12:29:27 UTC

This archive was generated by hypermail 2.3.1 : Friday, 3 May 2013 12:29:32 UTC