W3C home > Mailing lists > Public > public-gld-comments@w3.org > April 2013

Re: dataset and Dataset

From: Richard Cyganiak <richard@cyganiak.de>
Date: Sat, 6 Apr 2013 19:34:36 +0100
Cc: "'Bill Roberts'" <bill@swirrl.com>, <public-gld-comments@w3.org>
Message-Id: <F133FDEF-6ABB-46F1-B18B-973C6E37ED5A@cyganiak.de>
To: "Makx Dekkers" <makx@makxdekkers.com>
On 6 Apr 2013, at 19:11, "Makx Dekkers" <makx@makxdekkers.com> wrote:
>> AFAIK, DCMI namespaces contain *either* classes *or* properties but
>> not both. So the policy seems to address a case that will never occur
>> anyway. DCAT has a single namespace that contains both classes and
>> properties, so the concerns are different.
> 
> No, the concern was the same. For example, the class that is the range
> of dcterms:language could have been called dcterms:Language, but it is
> dcterms:LinguisticSystem precisely to avoid that confusion. 

So the policy is: If a class and property name would only differ in the case of the initial letter, then replace one with an awkward synonym. :-)

This is a somewhat common approach, unfortunately. FOAF does it as well (foaf:Image vs foaf:img, foaf:OnlineAccount vs foaf:account).

My personal view is that I'd rather resolve conflicts by paying careful attention to the “classes uppercase, properties lowercase” rule. To me, that seems easier for everyone than remembering two terms for what's really the same thing.

(Also, as I said elsewhere in the thread, outside of the SemWeb world, having a field or column or attribute "language" that points to a class or table or element "Language" wouldn't raise a single eyebrow.)

Best,
Richard
Received on Saturday, 6 April 2013 18:35:01 UTC

This archive was generated by hypermail 2.3.1 : Saturday, 6 April 2013 18:35:02 UTC