Re: audiobook options in objects

Hi Karen,

On Sat, Feb 9, 2013 at 5:49 PM, Karen Coyle <kcoyle@kcoyle.net> wrote:
> Niklas, I hope you don't mind if I take advantage of your knowledge for
> another round of questions...

Ask away and I'll do my best. :) I should note that I am more
acquainted with RDF patterns in general than with specific schema.org
usage. But given the information I've gathered on how schema.org
applies RDF thinking, and also where it refrains from certain
patterns, I think I've got a grasp of the situation.

(Note that I use the words "class" and "type" interchangeably below. I
believe that schema.org also uses "schemas" to mean classes, which is
how I interpret your first question below.)

> 1) is it acceptable practice to define a term in more than one schema? Or is
> it preferable to define your "thing" as being subordinate to multiple
> schemas? (I have looked and cannot find any examples of terms being
> "re-defined" in separate sub-classes.)

It depends. From a logical standpoint, it would be perfectly fine to
describe a thing as multiple types (as long as they're not disjoint),
and to use properties whose domain is either of those classes. (In
fact, the rdfs:domain semantics does imply that a thing is of a
specific type as long as it is described with a property having that
type as its domain.)

But I don't think this is the schema.org common practice. Its use of
domain (and range) is more loose (and hence it doesn't use the
rdfs:domain property, but schema:domain, which is more "annotational"
if you will). And if a property does belong to two schema.org classes
for which there is no usable base class, they do prescribe declaring
it to have multiple domains. Examples of that include e.g.:

    schema:actor
      schema:domain schema:Movie, schema:TVEpisode, schema:TVSeries .

    schema:address
      schema:domain schema:Organization, schema:Person, schema:Place .

(From a total of 114 occurrences of multiple domains.)

So to answer your question: I'd say yes to "define a term in more than
one schema", with the modification that you do not actually
(re-)define the term. Instead you describe it as having multiple
domains. It is still the same term, with the same meaning. This is not
obvious from reading the individual pages of each type at schema.org
though. You need to dig into the representation of the datamodel
(available in microdata and RDFa form from [1]). (I think this
documentation problem will be solved when schema.org starts publishing
pages for each property as well.)

.. That is not to say that multiple types aren't useful. If you mix
vocabularies together, it is quite applicable. As is "sugaring" your
data with multiple types and properties for different consumers (in
case there hasn't yet been declared a formal relationship between the
vocabularies, or they don't use RDFS/OWL inference (nor the simpler
subset of rules called RDFa vocabulary expansion)).

(Note also that if a usable base class is defined later on, it would
be perfectly ok to express this then. I.e. to insert it into the type
hierarchy, and to change the multiple statements of schema:domain on
properties that belong to the new base into one statement linking to
that one.)

> 2) if I were to define a term that had been previously used (like "isbn")
> but give it a different definition, that would be an error, right? and would
> not be allowed?

Yes, a term with a unique URI stands for one property and cannot be
redefined. If it is, it changes meaning; if there are multiple
conflicting statements about it, it is illogical. But I don't think
this is a problem here, since as per above a property isn't redefined,
but instead linked to multiple domains.

Cheers,
Niklas

[1]: http://schema.org/docs/datamodel.html


> kc
>
>
>
> On 2/9/13 2:56 AM, Niklas Lindström wrote:
>>
>> Hi Karen,
>>
>> On Sat, Feb 9, 2013 at 4:57 AM, Karen Coyle <kcoyle@kcoyle.net> wrote:
>>>
>>> Great, owen. I'll work on proposal 2 since that seems to be the mostly
>>> likely one in my set. I'm not clear about repetition of property names in
>>> schema.org -- is it legit to have two schemas that both have a property
>>> "isbn"? I believe this was discussed either here or on the public-vocab
>>> list, and I was left with the impression that the URI pattern is
>>>
>>> http://schema.org/[property]
>>
>>
>> Yes, that's correct. (So the URI for "isbn" is
>> <http://schema.org/isbn>. DanBri has said that those shall eventually
>> resolve as well.)
>>
>>> and not
>>>
>>> http://schema.org/Book/[property]
>>> http://schema.org/Audiobook/[property]
>>
>>
>> Indeed. As Jeff also replied, this pattern is for uncontrolled
>> user-derived classes.
>>
>>> Can anyone confirm? Because that would knock out a couple of the options
>>> that are currently on the page.
>>
>>
>> Schema.org properties can belong to more than one class (have more
>> than one domain), which don't have to be subclasses of one another.
>>
>> Quoting [1]:
>>
>> [[[
>> 1. each property may have one or more types as its domains. The
>> property may be used for instances of any of these types.
>> 2. each property may have one or more types as its ranges. The
>> value(s) of the property should be instances of at least one of these
>> types.
>>
>> The decision to allow multiple domains and ranges was purely
>> pragmatic. While the computational properties of systems with a single
>> domain and range are easier to understand, in practice, this forces
>> the creation of a lot of artifical types, which are there purely to
>> act as the domain/range of some properties.
>> ]]]
>>
>> Thus I think all is good with your proposals.
>>
>> Cheers,
>> Niklas
>>
>> [1]: http://schema.org/docs/datamodel.html
>>
>>
>>> kc
>>>
>>>
>>> On 2/8/13 3:46 PM, Owen Stephens wrote:
>>>>
>>>>
>>>> OK - I've given markup a go for Proposal 5 - not done this before so
>>>> likely be errors/misconceptions on my part so anyone should feel free to
>>>> chip in and improve.
>>>> http://www.w3.org/community/schemabibex/wiki/Object_Types#Proposal_5
>>>> Owen
>>>>
>>>> Owen Stephens
>>>> Owen Stephens Consulting
>>>> Web: http://www.ostephens.com
>>>> Email: owen@ostephens.com
>>>> Telephone: 0121 288 6936
>>>>
>>>> On 8 Feb 2013, at 18:01, Karen Coyle <kcoyle@KCOYLE.NET> wrote:
>>>>
>>>>> So many choices! :-) Should the next step be to try to encode a few
>>>>> items
>>>>> and see how it works?
>>>>>
>>>>> Also, the creativeWork "encoding" is *supposed* to be of type
>>>>> MediaObject. Does that make a difference?
>>>>>
>>>>> kc
>>>>>
>>>>> On 2/8/13 7:46 AM, Owen Stephens wrote:
>>>>>>
>>>>>>
>>>>>> Hi Karen,
>>>>>>
>>>>>> I'm inclined to ignore that definition of Media Object :) There is no
>>>>>> reason why an Audio-book shouldn't be embedded in a webpage so I
>>>>>> wouldn't
>>>>>> want to exclude this either)
>>>>>>
>>>>>> I think you could make use of the Creative Work property "encoding"
>>>>>> rather than making it a type of Media Object - which allows some split
>>>>>> of
>>>>>> content and carrier? (not wishing to resurrect that discussion!)
>>>>>>
>>>>>> This would allow:
>>>>>>
>>>>>> Proposal
>>>>>> Thing > CreativeWork > Audiobook
>>>>>> Thing > CreativeWork > Book > Audiobook
>>>>>>
>>>>>> New properties for audiobook type
>>>>>>
>>>>>>          • readBy -- expected type: Person
>>>>>>
>>>>>> Properties likely to be used from Book
>>>>>>
>>>>>>          • bookEdition (for abridgement note)
>>>>>>          • isbn
>>>>>>
>>>>>> With a separate Media Object to be created and referenced using
>>>>>> 'encoding' property from CreativeWork?
>>>>>>
>>>>>> I've added as Proposal 4
>>>>>>
>>>>>> Owen
>>>>>>
>>>>>> Owen Stephens
>>>>>> Owen Stephens Consulting
>>>>>> Web: http://www.ostephens.com
>>>>>> Email: owen@ostephens.com
>>>>>> Telephone: 0121 288 6936
>>>>>>
>>>>>> On 8 Feb 2013, at 15:23, Karen Coyle <kcoyle@kcoyle.net> wrote:
>>>>>>
>>>>>>> I wrote up three options for audiobook:
>>>>>>>
>>>>>>> http://www.w3.org/community/schemabibex/wiki/Object_Types#AudioBook
>>>>>>>
>>>>>>> A key question here is whether we can re-use MediaObject, since it is
>>>>>>> defined as being media embedded in a web page.
>>>>>>>
>>>>>>> kc
>>>>>>> --
>>>>>>> Karen Coyle
>>>>>>> kcoyle@kcoyle.net http://kcoyle.net
>>>>>>> ph: 1-510-540-7596
>>>>>>> m: 1-510-435-8234
>>>>>>> skype: kcoylenet
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>> --
>>>>> Karen Coyle
>>>>> kcoyle@kcoyle.net http://kcoyle.net
>>>>> ph: 1-510-540-7596
>>>>> m: 1-510-435-8234
>>>>> skype: kcoylenet
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>> --
>>> Karen Coyle
>>> kcoyle@kcoyle.net http://kcoyle.net
>>> ph: 1-510-540-7596
>>> m: 1-510-435-8234
>>> skype: kcoylenet
>>>
>>
>
> --
> Karen Coyle
> kcoyle@kcoyle.net http://kcoyle.net
> ph: 1-510-540-7596
> m: 1-510-435-8234
> skype: kcoylenet

Received on Sunday, 10 February 2013 12:32:48 UTC