Dear all,
I am frequently asked whether GoodRelations provides classes and
properties for describing the product or services included in an offer etc.
There seems to be a bit of confusion. GoodRelations provides two things:
1. A carefully designed set of classes, properties, and individuals for
describing the offer and demand relationships between a business entity
and a product.
2. A top-level ontology for specific products and services ontologies
for describing functional aspects of products and services included in
offers for sale, lease, repair, etc., i.e.
- gr:ProductOrService as a superclass and
- gr:quantitatativeProductOrServiceProperty, gr:
qualitativeProductOrServiceProperty, gr:datatypeProductOrServiceProperty
as superproperties for product features, and
- a bit more.
In "minimal" mode, you can use just part 1 of GoodRelations; it still
buys you a lot, because you can combine "semantic" search with fulltext
search on a much smaller subset
See here:
http://www.ebusiness-unibw.org/wiki/GoodRelationsQueries (draft)
Compare a Google search for "camcorder" with a faceted search for all
GoodRelations offers (gr:Offering) that include at least one
gr:ProductOrServicesSomeInstancesPlaceholder, to which an rdfs:label or
rdfs:comment is attached that contains the string "camcorder". You will
- search a much, much smaller text corpus (maybe 80 chars instead of
Terabytes of text)
- narrow your search to English content using the RDF language tag (and
expand it to other languages using Wikipedia etc.)
Detailed properties for describing the object or service (a camcorder, a
car, an apartment,...) are being provided by GoodRelations-compliant
ontologies for individual vertical industries, like eClassOWL,
freeClass, Consumer Electronics Ontology (CEO), etc.
See here:
http://www.ebusiness-unibw.org/wiki/GoodRelations#Compatible_Vocabularies_for_Products_and_ServicesSuch can be provided and are being prepared by many interest groups in
relevant markets.
Also, you can use dbPedia URIs and/or turn proprietary catalog
hierarchies into GoodRelations-compliant ontologies for describing the
product in more granularity.
There will be recipes for those two alternatives at
http://www.ebusiness-unibw.org/wiki/GoodRelationsComprehensiveDBpediaand
http://www.ebusiness-unibw.org/wiki/GoodRelationsComprehensiveCatalogGroupsThe only important thing is that everybody uses the minimal top-level
ontology part for product types and product features, as described here:
http://www.ebusiness-unibw.org/wiki/GoodRelations#Creating_New_Vocabularies_for_Products_and_ServicesIdeally, there will be one or just a few dominating ontologies for
product types, at least in a given domain. But you should expect a few
hundred in reality, and real business matchmaking on the Web of Linked
Data will require
- a sophisticated,
- iterative (find out how your types of interest are described - popular
properties etc.), and
- hybrid (combine structural/semantic and text/HLT/Regex) search
strategy - for a sketch, see
http://www.ebusiness-unibw.org/wiki/GoodRelationsQueriesAny shop will increase its visibility on the Giant Graph of Commerce
Data if he/she
- adds more granularity and
- chooses popular ontologies for the given vertical industry instead of
just publishing a proprietary vocabulary, even if that was GoodRelations
compatible.
This will be the real challenge for future Search Engine Optimization
using GoodRelations and RDFa. And there will be a trade-off decision
between the effort and the impact, depending on the quality of the
source data.
Many shops do currently just have a textual description of their
products in their databases. We cannot force them to lift all that to a
fully structured representation in one huge step, because they simply
can't do that. But they can gradually add more detail.
Also, I have high hopes in OpenCalais and other NLT/HLT products for
being able to lift minimal GoodRelations data to a more granular.
Again: GoodRelations supports a wide range of granularity - it really
depends on the technical ability of the owner of the data to provide
details.
Best wishes
Martin Hepp
--
--------------------------------------------------------------
martin hepp
e-business & web science research group
universitaet der bundeswehr muenchen
e-mail:
hepp at ebusiness-unibw.orgphone: +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 page:
http://purl.org/goodrelations/Resources for developers:
http://www.ebusiness-unibw.org/wiki/GoodRelationsWebcasts:
Overview -
http://www.heppnetz.de/projects/goodrelations/webcast/How-to -
http://vimeo.com/7583816Recipe for Yahoo SearchMonkey:
http://www.ebusiness-unibw.org/wiki/GoodRelations_and_Yahoo_SearchMonkeyTalk at the Semantic Technology Conference 2009:
"Semantic Web-based E-Commerce: The GoodRelations Ontology"
http://www.slideshare.net/mhepp/semantic-webbased-ecommerce-the-goodrelations-ontology-1535287Overview article on Semantic Universe:
http://www.semanticuniverse.com/articles-semantic-web-based-e-commerce-webmasters-get-ready.htmlTutorial materials:
ISWC 2009 Tutorial: The Web of Data for E-Commerce in Brief: A Hands-on Introduction to the GoodRelations Ontology, RDFa, and Yahoo! SearchMonkey
http://www.ebusiness-unibw.org/wiki/Web_of_Data_for_E-Commerce_Tutorial_ISWC2009
