W3C home > Mailing lists > Public > public-vocabs@w3.org > February 2012

Re: Schema.org property cardinality and use of plural (WAS Re: SoftwareApplication proposal for schema.org)

From: Jason Douglas <jasondouglas@google.com>
Date: Fri, 24 Feb 2012 13:33:35 -0800
Message-ID: <CAEiKvUCudx-P0Cfh5tjtH02W-gBWC=d=3=F9XHJtrttL6hdozw@mail.gmail.com>
To: Lin Clark <lin.w.clark@gmail.com>
Cc: Will Norris <will@willnorris.com>, "Evain, Jean-Pierre" <evain@ebu.ch>, Dan Brickley <danbri@danbri.org>, Adrian Giurca <giurca@tu-cottbus.de>, "public-vocabs@w3.org" <public-vocabs@w3.org>
On Fri, Feb 24, 2012 at 12:53 PM, Lin Clark <lin.w.clark@gmail.com> wrote:

> On Fri, Feb 24, 2012 at 3:25 PM, Will Norris <will@willnorris.com> wrote:
>
>
>> To address the original question, I think singular and plural naming of
>> properties is certainly one way (though not the only way) to imply the
>> expected cardinality of a particular property.  Yes, you get the somewhat
>> unusual mismatch of naming each individual instance with a plural name, but
>> you're always going to run into this certain cases.  If were you use
>> singular naming, you get weird results when you following the specified
>> JSON mapping:
>>
>> {
>>   "type": "http://schema.org/Person",
>>   "properties": {
>>     "child": [ { ... }, { ... } ]
>>   }
>> }
>>
>
> I would prefer to see the potentially confusing mismatch be on the JSON
> (consumer) side rather than the HTML (publisher) side.
>
> I say this because I think we can expect a higher degree of technical
> understanding from consumers who are processing the data than we can from
> Web content authors who are publishing the data. In order to move the
> mismatch from publishers to consumers, terms need to be singular
>

I think this is a very salient point.  Because you can only declare one
relationship at a time in markup, an itemprop effectively always seems
singular.  So when you have a plural property name:

<div itemscope itemtype="http://schema.org/Movie">
  <ul>
    <li itemprop="actors">actor1</li>
    <li itemprop="actors">actor2</li>
    <li itemprop="actors">actor3</li>
  </ul>
</div>

it seems kinda weird...

-jason


>
>
>> I think the same would apply to cardinality.  We provide guidance on
>> expected cardinality of properties, but always do the best we can with
>> whatever we get.
>>
>>
> I agree that it is good for large scale Schema.org consumers to do the
> best with whatever they get. However, that's more difficult for smaller
> scale projects. For example, I am working with the Drupal community to
> develop a tool that relies on HTML data to bring together content in
> sensible ways and I would like to rely on Schema.org for the
> vocabulary. The amount of available resource that we can put into this
> project is magnitudes smaller than the resources the search engines have
> available (I would assume, at least). Therefore, if there is mass confusion
> about how to use the terms, we're much less likely to be able to build upon
> it.
>
> -Lin
>
Received on Friday, 24 February 2012 21:34:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 06:48:59 GMT