Re: SemWeb terminology page

Hi

I still have problems with this proposal.
The point is that on Linked Data every dataset can be used as a reference for creating descriptions, be they corresponding to "bibliographic records" or "vocabularies" on the *traditional* library environment. To a great extent something like Worldcat can be used as a reference dataset for creating descriptions in other datasets. Linked Data-wise, there's not much functional difference between LCSH and Worldcat, then: they're just big bags of URIs you can re-use.
Note that using "dataset" alone was not the only option we discussed in Cologne: in fact we liked (and perhaps better) expressions like "organized dataset" or "reference dataset" (or "organized reference dataset" ;-) ) which capture this authority aspect better.

Antoine


>
> + 1 for Mark's categorization!
> Marcia
>
> On Dec 16, 2010, at 9:48, "Mark van Assem"<mark@cs.vu.nl>  wrote:
>
>> Hi,
>>
>> Dropping back in rather late in this thread:
>>
>> I would name
>>
>> - a set of bibliographic records     -->   "dataset"
>> - values records use (e.g. LCSH)     -->   "value vocabularies"
>> - properties and classes records use -->   "metadata element sets".
>>
>> In other words, I really dislike the suggestion of "dataset" for the
>> group of things that has LCSH etc. in it.
>>
>> If library folks would immediately identify "dataset" with "LCSH" then
>> it's fine, but then my question is what they would call a set of
>> bibliographic records.
>>
>> Mark.
>>
>> On 10/12/2010 2:04, Thomas Baker wrote:
>>> On Wed, Dec 08, 2010 at 08:01:34AM -0800, Karen Coyle wrote:
>>>> Note that Jenn Riley tackles semantic differences between library use
>>>> of terms and SemWeb use of terms in her (excellent!) slide presentation:
>>>>    http://www.dlib.indiana.edu/education/brownbags/fall2010/rdf/rdf.pdf
>>>> starting at slide 17. Some highlights:
>>>
>>> I'm on the road and this URL is timing out, so I cannot see
>>> the full document.  Just a few constructively intended
>>> reactions in response to the highlights...:
>>>
>>>> ?Subject?
>>>>
>>>> In libraries, what an information resource is about
>>>> In RDF, what a statement is about
>>>
>>> The "subject" of a triple, like the "subject" of
>>> a sentence, is of course what the statement is about.
>>>
>>> However, an RDF statement using, for example, the predicate
>>> dcterms:subject, says what the information resource is about
>>> -- no difference there.  I wouldn't want readers to think
>>> that RDF is somehow pushing people to think about "subject"
>>> in a completely alien way.  After all librarians, like all
>>> other users of natural-language grammar, form sentences about
>>> "subjects" just about every time they voice a thought :-)
>>>
>>>> "Class?
>>>>
>>>> In libraries, a classification scheme indicating the
>>>> general topic or area of knowledge covered by an
>>>> information resource
>>>> In RDF, a type or category that any type of object
>>>> or resource belongs to
>>>
>>> Hmm, would it not perhaps be more accurate to say:
>>>
>>>      In libraries, the general topic or area of knowledge
>>>      covered by an information resource as taken from (or
>>>      words to that effect) a classification scheme.
>>>
>>> Seen this way, the difference between the two is still there
>>> but is not quite as wide.
>>>
>>>> "Schema?
>>>>
>>>> XML Schema defines a set of elements intended to
>>>> be used together
>>>
>>> One could perhaps go one step further:
>>>
>>>      XML Schema defines a set of elements intended to
>>>      be used together in a specified document format.
>>>
>>>> RDF Schema defines classes and properties intended
>>>> to be used anywhere, alone or in combination
>>>
>>> Or more specifically:
>>>
>>>      RDF Schema defines classes and properties intended
>>>      to be used in RDF statements, either in isolation or
>>>      in the context of a set of statements.
>>>
>>> ...though that is perhaps too wordy.
>>>
>>> Tom
>>>
>>

Received on Thursday, 16 December 2010 14:57:57 UTC