Re: Another example of Wikidata + schema.org for type enumerations

On 25 Feb 2014, at 03:37, Young,Jeff (OR) <jyoung@oclc.org> wrote:

> RDFa Core 1.1 is a "W3C Recommendation", whereas Microdata is a "W3C Working Group Note".
> 
> http://www.w3.org/TR/rdfa-syntax/
> http://www.w3.org/TR/microdata/

Let's not open a thread on the history and politics of Microdata vs. RDFa. Fact is that we have to take both specs pretty much as given for the coming years, and the current situation between the camps is, IMO, okay to work with. We chose to solve the additional type problem in schema.org instead of at the Microdata spec level (even though it is mostly -- yet not solely -- useful for Microdata), because

1. That was much faster than waiting for a new Microdata spec.
2. A proper handling at the Microdata spec level is non-trivial, since a notion of a main type is needed for determining the possible properties, and their semantics. Since properties are bound to types in Microdata, we could have the same property for type 1 and type 2 with a different meaning. Then, which definition is to be used for a multi-typed entity? The additionalType property pattern solves that problem, since we always have a main type but are still able to expose additional type information.

Of course, schema.org currently uses global properties, but this is not more than a mere convention at this point, and I have already argued quite a few times that I think we may have to break that convention if the vocabulary grows beyond a certain size. Otherwise, we will run out of catchy property names.

Best wishes / Mit freundlichen Grüßen

Martin Hepp

-------------------------------------------------------
martin hepp
e-business & web science research group
universitaet der bundeswehr muenchen

e-mail:  martin.hepp@unibw.de
phone:   +49-(0)89-6004-4217
fax:     +49-(0)89-6004-4620
www:     http://www.unibw.de/ebusiness/ (group)
         http://www.heppnetz.de/ (personal)
skype:   mfhepp 
twitter: mfhepp

Check out GoodRelations for E-Commerce on the Web of Linked Data!
=================================================================
* Project Main Page: http://purl.org/goodrelations/




> 
> The key difference is obvious, right?
> 
> On Feb 24, 2014, at 9:32 PM, "Thad Guidry" <thadguidry@gmail.com> wrote:
> 
>> I disagree and do not think email documentation is the way forward for us.
>> 
>> We should not have to TELL Jarno this... we should have decent enough documentation / annotations / explains within Schema.org that make this clearer than mud.
>> 
>> We can do better.  I am sure of it..... thinking...
>> 
>> 
>> 
>> On Mon, Feb 24, 2014 at 8:21 PM, Stéphane Corlosquet <scorlosquet@gmail.com> wrote:
>> 
>> On Mon, Feb 24, 2014 at 8:44 PM, Jarno van Driel <jarno@quantumspork.nl> wrote:
>> And which also is confusing in the case of multiple type entities in Microdata.
>> 
>> I can imagine folks will write something like this:
>> 
>> <span itemscope itemtype="http://schema.org/Product">
>>     <link itemprop="additionalType" href="http://schema.org/Service">
>>     ...
>> </span>
>> 
>> as opposed to:
>> 
>> <span itemscope itemtype="http://schema.org/Product http://schema.org/Service">
>>     ...
>> </span>
>> 
>> Or is this something that should be accepted as correct markup?
>> 
>> They are both correct (if you assume that additionalType is the same as a regular type and your tooling can merge them). To make it easier to remember that @href and @src should only include one value, remember that these attributes are HTML attributes, and therefore any syntax built on top has to follow the HTML rules for these attributes. If you think that these attributes have to be interpreted and rendered in a browser, you definitely cannot include multiple URIs or things will break (broken links and broken images).
>> 
>> Steph.
>>  
>> 
>> 
>> On Tue, Feb 25, 2014 at 2:39 AM, Jarno van Driel <jarno@quantumspork.nl> wrote:
>> Well for me the confusement started with a remark of GuHa: "additionalType == typeOf" (http://lists.w3.org/Archives/Public/public-vocabs/2013Oct/0136.html).
>> 
>> Which got me to think that in case of additionalType one could write:
>> <link itemprop="additionalType" href="http://schema.org/Type1 http://schema.org/Type2">
>> 
>> Although Stéphane's remark: "href can only include one single URI" and Martin's remark: "the type in here is a property value" do make perfect sense from an HTML perspective. 
>> 
>> Now I looked at Dan's link to http://www.w3.org/TR/rdfa-syntax/#A-href and I've also looked it up in the Microdata specifications (http://www.w3.org/TR/2013/NOTE-microdata-20131029/#values) and one could argue that they do indicate a single URI. All be a bit technocratic. So IMO I think it would be a good thing it schema.org could explain this a bit more 'readable'.
>> 
>> 
>> On Tue, Feb 25, 2014 at 2:24 AM, Thad Guidry <thadguidry@gmail.com> wrote:
>> This is probably going to be a FAQ question over and over and over...so..
>> 
>> We should probably annotate when something takes multiple values within the schema somehow... hmmm.... something like... "only single value allowed"  or  "doesn't support multiple values".
>> 
>> Or is there already a hard and fast rule here in the schema... that only Types can take multiple values ?
>> 
>> Thoughts ?
>> 
>> -- 
>> -Thad
>> +ThadGuidry
>> Thad on LinkedIn
>> 
>> 
>> 
>> 
>> 
>> -- 
>> Steph.
>> 
>> 
>> 
>> -- 
>> -Thad
>> +ThadGuidry
>> Thad on LinkedIn

Received on Tuesday, 25 February 2014 10:25:07 UTC