W3C home > Mailing lists > Public > public-vocabs@w3.org > January 2014

Re: makesOffer should accept Service

From: Martin Hepp <martin.hepp@ebusiness-unibw.org>
Date: Tue, 7 Jan 2014 17:08:23 +0100
Cc: Dan Scott <dan@coffeecode.net>, Tyler Shuster <tyler.herrshuster@gmail.com>, SchemaDot Org <public-vocabs@w3.org>
Message-Id: <1884F324-A8E2-427E-BE17-3C6961F3042E@ebusiness-unibw.org>
To: Jarno van Driel <jarno@quantumspork.nl>
One reason why this proposal seems a bit more tempting in the schema.org context than in the general Semantic Web / Linked Data context is that schema.org in Microdata markup often does not assign global identifiers to objects (via itemid), so it is not immediately visible that e.g. three offers refer to the same product.

But the distinction is really valuable, because it allows a search engine to e.g. augment product information with data from additional sources (e.g. product features) without overriding offer data (e.g. the dealer price with the recommended retail price).

On Jan 7, 2014, at 8:52 AM, Jarno van Driel wrote:

> Thanks Dan for giving your interpretation. Now I hope you don't mind but your response raises some new questions to me.
> > 'The question "What is being offered?" is answered by the itemOffered property'...
> I understand that this is where triples come into play and that with it you can map relationships between entities but what confuses me is, when parsing markup like this through the different tools out there I get an array of types back which share the same properties. Doesn't this then imply there also is a relationship (even without the notation of itemOffered)?
> If not, then when extracting the data of that multi-type entity ('Offer Service' - to stay with the example) don't things go wrong? 
> e.g. an Offer has a price property while Service doesn't. Doesn't this then give back back wrong/invalid data? 
> On Tue, Jan 7, 2014 at 4:20 AM, Dan Scott <dan@coffeecode.net> wrote:
> On Mon, Jan 6, 2014 at 11:43 AM, Jarno van Driel <jarno@quantumspork.nl> wrote:
> Somebody asked me yesterday why itemtype="http://schema.org/Product http://schema.org/Service"> has to be marked up this way just to be able to add an offer/Offer. He proposed to mark it up like itemtype="http://schema.org/Offer http://schema.org/Service">. Adding the Offer as a second type of Service and thus skipping Product all together.
> In all honesty I couldn't come up with a reason why this would be a wrong notation. So I was curious, does anybody think this is valid markup and if not, why not?
> My interpretation is that you would describing something that is both an Offer and a Service--which is subtly but significantly different from describing something that is an Offer to provide a Service. The question "What is being offered?" is answered by the itemOffered property, which links an Offer to a Product. In the proposed markup, there's nothing for that property to point at, so a consumer of the markup would likely consider it a dead end.
> Of course, given enough encounters with this markup in the wild and/or a large enough customer insisting on the importance of this markup, it's possible that schema.org consumers would make a special case when they parse multi-typed items that include Offer as one of the types.

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

e-mail:  hepp@ebusiness-unibw.org
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/
Received on Tuesday, 7 January 2014 16:08:49 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:49:20 UTC