- From: Martin Hepp <martin.hepp@ebusiness-unibw.org>
- Date: Sat, 28 Jul 2012 23:51:52 +0200
- To: Kenley Lamaute <kenleyl@microsoft.com>
- Cc: Charlie Jiang <chjiang@microsoft.com>, Dan Brickley <danbri@google.com>, Joshua J Wulf <jwulf@redhat.com>, "public-vocabs@w3.org" <public-vocabs@w3.org>
I agree that isVariantOf (http://purl.org/goodrelations/v1#isVariantOf) is more generic than the proposed schema:productFamilyID. However, the question is whether we really need two properties for capturing the subtle difference. Again, I have no objections in adding this second property if it is really needed to be more specific than the upcoming schema:isVariantOf. But I personally am inclined to believe that, at Web scale, we can live with one property for the moment. Also I would like to stress again that Volkswagen is already heavily using gr:isVariantOf in their Car Options Ontology and related data. Best Martin On Jul 28, 2012, at 10:48 PM, Kenley Lamaute wrote: > Hmmm, yes there is a semantic difference between "variant" (could be used more broadly) and "family" (implies a close relationship). > > -----Original Message----- > From: Charlie Jiang > Sent: Saturday, July 28, 2012 10:11 AM > To: Martin Hepp; Kenley Lamaute; Dan Brickley > Cc: Joshua J Wulf; public-vocabs@w3.org > Subject: RE: currentModel attribute for /Product > > I wonder if there is subtle difference in English here. "isVariantOf" can mean broader than "productFamilyID". The former implies containment and also sometime branch-out siblings. The latter is only for containment. Am I right? > > -----Original Message----- > From: Martin Hepp [mailto:martin.hepp@ebusiness-unibw.org] > Sent: Saturday, July 28, 2012 2:30 AM > To: Kenley Lamaute; Dan Brickley > Cc: Joshua J Wulf; public-vocabs@w3.org > Subject: Re: currentModel attribute for /Product > > Dear Kenley, Dan: > > Apologies for the delay in checking the dependencies to the ongoing GoodRelations integration. > > As for adding > > releaseDate > > to > http://schema.org/Product: > > That is perfectly fine and should be done. We will also add this property to GoodRelations in its original namespace shortly. > > As for the other property, > > productFamilyID: > > After the GoodRelations addition, there will be a property > > isVariantOf > > for the new item type > > http://schema.org/ProductModel > > which basically caters the use-case of linking from a base model (e.g. a Volkswagen Golf family) to a more specific product model (e.g. a Volkswagen Golf Station Wagon 1.4 l). > > It will not harm to add the proposed productFamilyID property even if schema.org will allow more advanced patterns product model information soon. But the following markup would do the trick with just adding one property instead of two: > > <div itemscope itemtype="http://schema.org/Product" itemid="#model"> > <span itemprop="name">ACME Colorvision 123</span> > <span itemprop="description">The ACME Colorvision 123 is the leading-edge color TV from our company.</span> > Version<meta itemprop="releaseDate" content="2012-01-01">Jan 1, 2012</span> > <link itemprop="isVariantOf" itemscope itemtype="http://schema.org/ProductModel" itemid="URI of the base model" /> </div> > > This pattern would also be compatible with what Volkswagen is doing in their Car Options Ontology, http://purl.org/coo/ns. > > So as bottom line, it will be okay to add both properties, but adding releaseDate would be sufficient. > > As a side-note: After the GoodRelations addition, it will be possible (and cleaner), to distinguish between markup for products and markup for product datasheets, so using http://schema.org/ProductModel will be a bit better. > > Best wishes > > Martin Hepp > > > On Jul 27, 2012, at 7:18 AM, Kenley Lamaute wrote: > >> Good point. ProductFamilyURI and ReleaseDate seem like a winning combo. >> From: Joshua J Wulf [jwulf@redhat.com] >> Sent: Thursday, July 26, 2012 10:11 PM >> To: Kenley Lamaute >> Cc: public-vocabs@w3.org >> Subject: Re: currentModel attribute for /Product >> >> Yes, a URI is best, because otherwise there is no guaranteed way to avoid namespace collision. >> >> In that case, maybe ProductFamilyURI might be better? It could then additionally be dereferenceable, with no need to specify at this stage what that might return, but leaving the door open to future developments there... >> >> On 07/27/2012 02:52 PM, Kenley Lamaute wrote: >>> Yes any unique ID could do; however, we can certainly offer guidance. One simple approach could be to use a URL for the product family ID value, as a fully qualified URL is universally unique. >>> >>> Thoughts? >>> >>> From: Joshua J Wulf [mailto:jwulf@redhat.com] >>> Sent: Thursday, July 26, 2012 9:42 PM >>> To: Kenley Lamaute >>> Cc: public-vocabs@w3.org >>> Subject: Re: currentModel attribute for /Product >>> >>> There will be guidelines on how it is to be generated? >>> >>> On 07/27/2012 02:38 PM, Kenley Lamaute wrote: >>> Yes, the ID is to be universally unique. >>> >>> From: Joshua J Wulf [mailto:jwulf@redhat.com] >>> Sent: Thursday, July 26, 2012 9:25 PM >>> To: Kenley Lamaute >>> Cc: public-vocabs@w3.org >>> Subject: Re: currentModel attribute for /Product >>> >>> On 07/27/2012 02:16 PM, Kenley Lamaute wrote: >>> I've received direct feedback from several individuals on the Product item proposal. After vetting several options, The most elegant solution for reducing the complexity of requiring metadata updates to already published pages, is to introduce two new fields to 'Product': >>> >>> * releaseDate - the product release date >>> * productFamilyID - unique product family identifier reused across product versions. For example: <meta itemprop='productFamilyID' content='123-456-789'/>. >>> >>> This enables the scenario of a user searching for information about a product (without specifying a version), and enabling the search engine to favor results for the most recent product material. >>> >>> Example: >>> *Legacy Product Content* >>> Product = Windows Mobile 6.5 >>> Model = 6.5 >>> releaseDate=10/6/2009 >>> productFamilyID=123-456-789 >>> >>> >>> *Newer Product Content* >>> Product = Windows Phone 7 >>> Model =7 >>> releaseDate=10/21/2010 >>> productFamilyID=123-456-789 >>> >>> The example above shows how multiple products can be associated by a ProductFamilyID, so that anyone seeking "mobile sdk for Windows" can be offered content that applies to the latest technology first. >>> >>> If there are no objections, I'd like to revise the original proposal for adding 'currentModel' to product, and replace it with this solution of adding 'releaseDate' and 'productFamilyID' to the Product item. >>> >>> -Kenley >>> >>> Killer. >>> >>> Given that you can equate different Products by giving them the same productFamilyID, does the productFamilyID need to be a UUID? >>> >>> - Josh >>> >>> >> >> > > -------------------------------------------------------- > 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/ > > > > > > > > > -------------------------------------------------------- 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 Saturday, 28 July 2012 21:52:20 UTC