From martin.hepp at ebusiness-unibw.org Wed Jun 1 23:46:56 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Wed, 1 Jun 2011 23:46:56 +0200 Subject: [goodrelations] New GoodRelations Landing Page and Wiki Message-ID: <73144E7F-A75F-4729-AE29-73125F510DAC@ebusiness-unibw.org> Dear all: We just deployed a completely redesigned landing page for GoodRelations and the new community & developer's wiki: http://purl.org/goodrelations/ Any feedback will be very welcome! Note: This does not change any existing applications, data, or queries. It is just a visual relaunch ;-) In the next coming days, we will also deploy a much improved version of the GoodRelations specification in HTML. A big thanks to my team members Andreas Radinger, Alex Stolz, Uwe Stoll, and L?szl? T?r?k! Best wishes Martin Hepp -------------------------------------------------------- martin hepp e-business & web science research group universitaet der bundeswehr muenchen e-mail: hepp at 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 From richard.hancock at 3kbo.com Mon Jun 6 02:43:14 2011 From: richard.hancock at 3kbo.com (richard.hancock at 3kbo.com) Date: Sun, 5 Jun 2011 19:43:14 -0500 Subject: [goodrelations] Listing Products sold at a Location Message-ID: <17131adfd3c1da4424dfda82de678d81.squirrel@www.3kbo.com> Hi all, I am interested in finding out how to most efficiently geo locate products that have been listed for sale using GoodRelations. For example a recipe lists onions and tomatoes as ingredients. I'd like to find those ingredients close to my current location (using latitude and longitude). Some initial attempts using gr:offers left me thinking that for scenarios like this it would be simpler having a property e.g. :sells (shown below) which lists (summarizes) the types of products being sold, so that the SPARQL query does not need to traverse through all the gr:Offering instances to reach the products to see if they are the ingredients I am interested in. The basic pattern I'm thing of is shown in the fragment below,using common names for the products that map easily to Wikipedia and hence to productontology.org. Is there an existing GoodRelations property like :sells that would be better to use? Are there other, better patterns for getting a quick listing of the products being sold at a particular location? @prefix gr: . @prefix pto: . :CometBayFarm a gr:BusinessEntity ; gr:offers :OnionsOffering , :TomatoesOffering ; :sells :Tomatoes , :Onions ; vcard:adr :address ; vcard:geo :location . :sells rdf:type owl:ObjectProperty ; rdfs:domain gr:BusinessEntity ; rdfs:range gr:ProductOrService . :Tomatoes a pto:Tomato , gr:ProductOrServicesSomeInstancesPlaceholder; rdfs:comment "Tomatoes organically grown at Comet Bay Farm"@en ; rdfs:label "Tomatoes"@en . Full example list: http://www.3kbo.com/examples/goodrelations/comet-bay-farm/goodrelations.n3 http://www.3kbo.com/examples/goodrelations/comet-bay-farm/goodrelations.rdf Cheers, Richard From martin.hepp at ebusiness-unibw.org Mon Jun 6 08:31:00 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Mon, 6 Jun 2011 08:31:00 +0200 Subject: [goodrelations] GoodRelations & Schema.org Message-ID: <1790B27B-6AB1-412A-8F22-F777CE93DA24@ebusiness-unibw.org> Dear all: You may have seen that Google, Yahoo, and Bing have just released a single Microdata-based data schema for adding rich data to Web sites. I am in the process of writing a more detailed analysis and recommendations, but just as a quick sketch: 1. Both Google and Yahoo have stressed that they continue to accept GoodRelations. 2. Surprisingly, Bing has announced on the very day (!) on which they joined the "rich snippets movemenent", that they will support GoodRelations in the future: http://onlinehelp.microsoft.com/en-us/bing/hh207242.aspx 3. Currently, GoodRelations in RDFa syntax is the only way of feeding Google and Yahoo with W3C-valid markup (XHTML+RDFa1.0). All other formats, including HTML5 Microdata, do, at this point, NOT validate. As for their motivation, I think that Google had hoped for a greater adoption of rich snippet data in either hproduct/hlisting, GoodRelations, or Microdata syntax since their announcement in November, 2010. They may also have the hope of simplifying their task of processing rich data by imposing more constraints on the data patterns by their sheer market power. However, I am confident that Schema.org will NOT fly, for the following reasons: 1. The new schema.org Microdata format for products is basically the very same proposed by Google in parallel to their support of GoodRelations in November 2, 2011. I do not have data at hand, but I am pretty confident that more people used either hProduct/hListing or GoodRelations in RDFa than the Google Microdata syntax. I am totally unconvinced that, with a Google market share of up to 90 % in most countries, the bundling of forces with Bing and Yahoo will fundamentally change the adoption. If Webmasters were not enthusiastic about the Google Microdata syntax described at http://www.google.com/support/webmasters/bin/answer.py?hl=en&answer=146750 why should they suddenly adopt it eagerly from http://schema.org/Product ? Nothing else has changed. 2. The main bottlenecks for the bit slow adoption of rich data in Web content in general are: a) The fact that Google tolerates, but not officially mandates RDFa in snippet-style [1], which would break up the complexity caused by mixing the organization of visible content and data structures. Also, RDFa in snippet-style makes it a lot easier to write extension modules for popular shop software. b) The fact that Google white-lists rich snippets for pages overly slow, so it takes months for many pages until they see the effect of rich data in the form of rich snippets. c) Data modeling is inherently more complex that "hobby" HTML development, i.e. writing a bit of markup and then validating whether it looks as you like it. Modern Web technology is difficult - Javascript, AJAX, SQL, HTML5 - all that has raised the bar for participation. The new Microdata syntax proposal does not take away this complexity. In fact, adopting the new markup manually is way more difficult than creating a GoodRelations snippet using the Snippet Generator at http://www.ebusiness-unibw.org/tools/grsnippetgen/ There is more to say on this, but at this point it is definitely save to stick to GoodRelations in RDFa syntax. Best wishes Martin [1] Hepp, Martin; Garc?a, Roberto; Radinger, Andreas: RDF2RDFa: Turning RDF into Snippets for Copy-and-Paste, Technical Report TR-2009-01, 2009. PDF at http://www.heppnetz.de/files/RDF2RDFa-TR.pdf From martin.hepp at ebusiness-unibw.org Mon Jun 6 17:26:20 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Mon, 6 Jun 2011 17:26:20 +0200 Subject: [goodrelations] GoodRelations in Google Rich Snippets: gr:validThrough now correctly handled & mandatory Message-ID: <4EFA1E2B-83D0-495C-B0AE-74EEE7EB5D9D@ebusiness-unibw.org> Dear all: I just found out that Google is now correctly using gr:validThrough for price information: You MUST indicate a validity date that has not yet lapsed for gr:UnitPriceSpecification in order to be accepted as a valid price.

Canon Offer

Average rating: 4.4, based on 89 reviews
... USD $899.00 (Valid from : December 30, 2009, through --->>>>> December 30, 2014)
VAT included

This again shows that Google is increasing its support of GoodRelations. Best wishes Martin Hepp From ml-ktk at netlabs.org Mon Jun 6 18:36:40 2011 From: ml-ktk at netlabs.org (Adrian Gschwend) Date: Mon, 06 Jun 2011 18:36:40 +0200 Subject: [goodrelations] GoodRelations & Schema.org In-Reply-To: <1790B27B-6AB1-412A-8F22-F777CE93DA24@ebusiness-unibw.org> References: <1790B27B-6AB1-412A-8F22-F777CE93DA24@ebusiness-unibw.org> Message-ID: <4DED0218.5080301@netlabs.org> On 06.06.11 08:31, Martin Hepp wrote: Hi Martin, > I am in the process of writing a more detailed analysis and > recommendations, but just as a quick sketch: excellent remarks, hope you can put them on a blog/page for linking to it. I made a post as well about the subject: http://strangelove.netlabs.org/2011/06/schema-org-not-too-impressive/ cu Adrian From martin.hepp at ebusiness-unibw.org Mon Jun 6 18:40:35 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Mon, 6 Jun 2011 18:40:35 +0200 Subject: [goodrelations] Using GoodRelations for Google AND Yahoo AND the Semantic Web Message-ID: <43CF3304-5614-4132-ACFC-6628C22745A4@ebusiness-unibw.org> Dear all: While updating the GoodRelations test-cases at http://www.heppnetz.de/rdfa4google/testcases.html which we normally use to monitor their validators, I found many examples of 1. using GoodRelations for both Yahoo and Google and 2. using extensions for specific verticals, e.g. http://purl.org/vso/ns.html for the auto industry, while still validating with Google AND Yahoo. This shows that, internally, they can handle way more of RDFa and GoodRelations than publicly advertised to developers. So when you want to publish rich data to please both search engines, mobile applications, and browser extensions, GoodRelations in RDFa is currently the best choice. The only fundamental difference between Google's and Yahoo's recommendation for using GoodRelations is the markup for review data; they use two different vocabularies. But thanks to RDFa's flexibility, you can still use a single block of markup:

Product Reviews:

Average: 4.5, lowest: 0, highest: 5 (total number of reviews: 45)
Best wishes Martin Hepp From martin.hepp at ebusiness-unibw.org Tue Jun 7 20:04:20 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Tue, 7 Jun 2011 20:04:20 +0200 Subject: [goodrelations] RWW Post on Schema.org References: <00CA67A9-C5B0-4B43-A679-E86F255A301C@ebusiness-unibw.org> Message-ID: <18146DCA-7D06-42F6-8009-C3CDE43F9A64@ebusiness-unibw.org> Dear all: I just posted a longer comment on www.readwriteweb.com, which may be relevant for the subscribers of this list, too: Best Martin Hepp Begin forwarded message: > http://www.readwriteweb.com/archives/is_schemaorg_really_a_google_land_grab.php#disqus_thread > > Dear Richard, > great post! > > First a minor correction: Google and Yahoo have confirmed that they will continue to support GoodRelations in RDFa for product and offer information. The schema.org markup is simply a bit more prominently featured way of feeding such data into Google and does not differ significantly from their 2010-11-02 microdata recipe. > > Second, interestingly, Bing announces on the very same page [1] both their schema.org recipes AND the indication that they will add GoodRelations support to their crawlers shortly. > > Given the amount of shop extensions [2] and the usage of GoodRelations in pretty sophisticated ways by large companies like BestBuy or Volkswagen [3], where the power of RDFa is needed, I am confident that all three search engines will stick to their support of GoodRelations. > > In fact, the simplification of the new microdata recipes is not as significant that it will really make a difference - in particular, if you compare it with the ease of GoodRelations snippets for copy-and-paste, created with the GoodRelations Snippet Generator [4] or the shop extensions for Magento or Joomla/Virtuemart. > > Also, while schema.org may work for the US search market, the deployment of Rich Snippets in other countries will require additional properties, e.g. for shipment costs, because you need this kind of details for complying with regional legislation. For instance, a German shop must indicate the shipment charges in any price overview listing. The microdata approach for schema.org requires a one-size-fits-all vocabulary, which means you either have to bother all developers in the world with more properties (e.g. for VAT, shipment options, warranties), or miss those features. > > So in a nutshell, I think that schema.org is first and most an attempt to foster the creating of simple, lightweight markup. For more sophisticated usages, GoodRelations in RDFa remains the better choice and will stay on Google's agenda. See the BestBuy and Volkswagen use-cases - neither application is doable with schema.org elements. > > From my experience with GoodRelations, the main reason for the bit slow adoption of rich markup by smaller sites is simply the fact that Google's manual whitelisting process for rich snippets - no matter whether based on hProduct/hListing, Microdata, or GoodRelations in RDFa - is extremely slow. This frustrates people. If you add markup that validates with the Google validator, you would expect a rich snippet to shop up in a few weeks or so. But up to today, site owners often wait months until they see the effect. > > If you do not believe that Google and Yahoo still heavily parse and process GoodRelations in RDFa, look at my suite of testcases that use very sophisticated GoodRelations patterns and still validate: > > http://www.heppnetz.de/rdfa4google/testcases.html > > I don't think that Google developed a Web-scale infrastructure for parsing GoodRelations in RDFa just for the dustbin. > > There may be a bit of dust covering this, and some wording on the schema.org site is misleading, but site-owners still have the choice to use either GoodRelations in RDFa or schema.org elements in Microdata syntax for e-commerce sites. > > Best wishes > > Martin Hepp > http://purl.org/goodrelations/ > > 1. http://onlinehelp.microsoft.com/en-us/bing/hh207242.aspx > 2. http://wiki.goodrelations-vocabulary.org/Shop_extensions > 3. http://www.volkswagen.co.uk/vocabularies/vvo/ns and http://www.volkswagen.co.uk/vocabularies/coo/ns > 4. http://www.ebusiness-unibw.org/tools/grsnippetgen/ > From tlc35us at yahoo.com Tue Jun 7 22:15:37 2011 From: tlc35us at yahoo.com (dan r) Date: Tue, 7 Jun 2011 13:15:37 -0700 (PDT) Subject: [goodrelations] Hello, questions about describing service area, sunset, and more Message-ID: <267198.69171.qm@web39405.mail.mud.yahoo.com> Dear All, My name is Daniel Reed. I work for a mobile service company in California, USA. When I'm not providing service for a customer I integrate our administrative work flow into a SaaS CRM. Nothing too technical just some HTML forms, CSS, and cut-and-paste Javascript. I'm glad that you have an online annotator because? when RDFa gets complex I'm not sure how to properly code it. ? There are a few things I'd like to change in the results: 1. How would the CA subdivision of US be stated here? ?
2. The business, stump removal, is classified as a tree service plus underground construction, therefore I give an estimate for the known portion of the service and the likely best case for the unknown portion. If the unforeseen or unexpected happens the customer is aware that the additional work is at the hourly rate. So I'd like to make them aware that I accept checks according to local customs ( gr:COD for cash or "local check" ) but not gr:CheckInAdvance since the actual price is known upon completion of service. I was hoping that if my terms weren't available the solution would be simple like gr:PaymentMethod xsd:comment"check", especially for gr:closes 3.The telephone callback hours are implied in our online survey 0800h-2000h. But the service hours are mainly daylight hours. So I'd like to express them as open: 0800h and close: sunset. Also, Saturday "by previous appointment only" ( I've noticed lately that many dentists are working Saturday "by appointment only").? I think of sunset as a relative term, but could it be a recurring pattern as mentioned in: http://ebusiness-unibw.org/pipermail/goodrelations/2011-May/000355.html ? In the same post, I don't understand, "wrap a SPARQL endpoint around?" . But would a link to a wikipedia page about the concept of sunset work? I'm trying to avoid multiple instances of gr:close for month and day-of-week all year long. ? Even more precisely would be "close: sunset or local ordinance whichever is earlier". I'm content to live without this precision but it might be interesting from a modeling perspective. If locksmiths and automotive roadside assistance companies were to adopt good relations en masse then it would be useful to have a model like "sunset" for "after dark" since their terms of service change at nightfall. 4.There are two common ways of specifying a service area. One is a radius around a given location and the other is to specify a potentially irregularly shaped area in this case specifically, a county. Therefore I have a need for greater geographical granularity. Google places allows me to specify a service area by listing three counties and two cities. Google then includes all points in between. So in keeping with this concept, perhaps five vcards similar to this: BEGIN:VCARD VERSION:3.0 N:Service Area 1 ORG:Stump King LLC ADR;TYPE=WORK:;;Santa Clara County;CA;United States of America END:VCARD (modified from wikipedia example) then the snippet might be:
?????
????????
???????????
??????????????
?????????????????
?????????????????
?????????????????
?????????????????
?????????????????
?????????????????
?????????????????
??????????????
???????????
<> ???????????
??????????????
?????????????????
?????????????????
?????????????????
??????????????
???????????
< ??????????? . ??????????? . ??????????? . ????????
?????
?The hard way, I imagine, would be to approximate the corners of the service area shaded by Google with lat/long coordinates. Currently, the service area is implied by listing many local telephone numbers and cities. 5. The business has seven telephone numbers each listed next to an appropriate city. Fine for people and current SEO, but what about machines? It seems reasonable to have a vcard with seven TEL properties. Do machines need a pairing of city and telephone number? If so, seven vcards with n,adr, and tel properties only? (Where adr is only the city name). This is without pretending to have seven stores or gr:hasPOS Thank you for your help, Daniel Reed -------------- next part -------------- An HTML attachment was scrubbed... URL: From martin.hepp at ebusiness-unibw.org Tue Jun 7 22:41:59 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Tue, 7 Jun 2011 22:41:59 +0200 Subject: [goodrelations] Hello, questions about describing service area, sunset, and more In-Reply-To: <267198.69171.qm@web39405.mail.mud.yahoo.com> References: <267198.69171.qm@web39405.mail.mud.yahoo.com> Message-ID: <8BBB9A50-BD96-448F-9721-04759663C06E@ebusiness-unibw.org> Hi Dan: On Jun 7, 2011, at 10:15 PM, dan r wrote: > Dear All, > My name is Daniel Reed. I work for a mobile service company in California, USA. When I'm not providing service for a customer I integrate our administrative work flow into a SaaS CRM. Nothing too technical just some HTML forms, CSS, and cut-and-paste Javascript. I'm glad that you have an online annotator because when RDFa gets complex I'm not sure how to properly code it. > There are a few things I'd like to change in the results: > > 1. How would the CA subdivision of US be stated here? >
The list is e.g. at http://en.wikipedia.org/wiki/ISO_3166-2:US California is "US-CA": >
> > 2. The business, stump removal, is classified as a tree service plus underground construction, therefore I give an estimate for the known portion of the service and the likely best case for the unknown portion. If the unforeseen or unexpected happens the customer is aware that the additional work is at the hourly rate. So I'd like to make them aware that I accept checks according to local customs ( gr:COD for cash or "local check" ) but not gr:CheckInAdvance since the actual price is known upon completion of service. I was hoping that if my terms weren't available the solution would be simple like gr:PaymentMethod xsd:comment"check", especially for gr:closes I would use gr:COD. For more specific payment options ("pay me in gold metals"), you can create a shop-specific instance of gr:PaymentMethod and define it in your main page (or any other page):
Gold (gr:PaymentMethod)
We accept pure gold as a form of payment
You could then say
> > 3.The telephone callback hours are implied in our online survey 0800h-2000h. But the service hours are mainly daylight hours. So I'd like to express them as open: 0800h and close: sunset. I would either use 0800-2000 or create multiple statements that are constrained using gr:validFrom and gr:validThrough - e.g., indicate individual opening hours for each month of the year. Computers need data; they cannot infer enough precision from a text like "sunset". What you can to, though, is adding additional text to the gr:OpeningHoursSpecification element using rdfs:comment. > Also, Saturday "by previous appointment only" ( I've noticed lately that many dentists are working Saturday "by appointment only"). I think of sunset as a relative term, but could it be a recurring pattern as mentioned in: > http://ebusiness-unibw.org/pipermail/goodrelations/2011-May/000355.html ? You could create a gr:OpeningHoursSpecification node for gr:Saturday, omit gr:opens and gr:closes and just attach your text using rdfs:label > In the same post, I don't understand, "wrap a SPARQL endpoint around?" . That means that you can expose a database or other dynamic data source as if it was an RDF database with all combinations materialized, while they are actually created only when a relevant query arrives. SPARQL is a W3C query language and interface specification for RDF data. > But would a link to a wikipedia page about the concept of sunset work? No, unfortunately. > I'm trying to avoid multiple instances of gr:close for month and day-of-week all year long. > Even more precisely would be "close: sunset or local ordinance whichever is earlier". I'm content to live without this precision but it might be interesting from a modeling perspective. If locksmiths and automotive roadside assistance companies were to adopt good relations en masse then it would be useful to have a model like "sunset" for "after dark" since their terms of service change at nightfall. The only way to model these terms in a form understandable for a machine is to compute those hours for any given day and create specific opening hour statements for each single day. But this would be a pretty sophisticated use-case. I would work with monthly approximations. > > 4.There are two common ways of specifying a service area. One is a radius around a given location and the other is to specify a potentially irregularly shaped area in this case specifically, a county. Therefore I have a need for greater geographical granularity. Google places allows me to specify a service area by listing three counties and two cities. Google then includes all points in between. So in keeping with this concept, perhaps five vcards similar to this: > BEGIN:VCARD > VERSION:3.0 > N:Service Area 1 > ORG:Stump King LLC > ADR;TYPE=WORK:;;Santa Clara County;CA;United States of America > END:VCARD > (modified from wikipedia example) > then the snippet might be: >
>
>
typeof="gr:LocationOfSalesOrServiceProvisioning"> >
>
typeof="vcard:Address"> >
>
>
>
>
>
>
>
>
> <> >
>
typeof="vcard:Address"> >
>
>
>
>
> < > . > . > . >
>
>
> > The hard way, I imagine, would be to approximate the corners of the service area shaded by Google with lat/long coordinates. Currently, the service area is implied by listing many local telephone numbers and cities. GoodRelations does currently not allow specifying arbitrary geographical areas. You can only go down to ISO 3166 granularity. We may extend this in the future. > > 5. The business has seven telephone numbers each listed next to an appropriate city. Fine for people and current SEO, but what about machines? It seems reasonable to have a vcard with seven TEL properties. Do machines need a pairing of city and telephone number? If so, seven vcards with n,adr, and tel properties only? (Where adr is only the city name). This is without pretending to have seven stores or gr:hasPOS By far the best modeling is to create individual instances of gr:Location (new name for gr:LocationOfSalesOrServiceProvisioning, will become the preferred name tomorrow) for each service location and attach the phone numbers to those locations. So seven gr:Location with seven vcards, seven tel properties. > Thank you for your help, You are very welcome! Martin > Daniel Reed > _______________________________________________ > goodrelations mailing list > goodrelations at ebusiness-unibw.org > http://ebusiness-unibw.org/cgi-bin/mailman/listinfo/goodrelations From martin.hepp at ebusiness-unibw.org Thu Jun 9 00:44:22 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Thu, 9 Jun 2011 00:44:22 +0200 Subject: [goodrelations] GoodRelations: Site relaunch, service update Message-ID: <7D3529E0-0067-4200-A65D-03FA8DE2DA94@ebusiness-unibw.org> Dear all: We just deployed the new GoodRelations Web site and a GoodRelations Service Update at http://purl.org/goodrelations/ Please refresh your caches and repositories. The relaunch is a major step to a more developer-friendly environment. Here is a summary of the changes: - Social Web functionality for each individual GoodRelations element - you can easily share and refer friends to each and every element directly - Tight integration with popular developer forums, like Quora, StackOverflow, answers.semanticweb.com, so that questions related to parts of GoodRelations can be shared with relevant experts with ease - Cross-references to external resources related to an element (e.g. a link to Wikipedia pages when a property value is from a standard code) - Examples in Turtle, RDF/XML, RDFa, and SPARQL for each element (this is currently work in progress) - New screen-design and Wiki The Service Update 2011-04-01 also brings a few modifications to the vocabulary. They are implemented in a backwards-compatible fashion and approved by Yahoo and Google, so their parsers will tolerate the old and the new versions. gr:ActualProductOrServiceInstance --> gr:Individual gr:ProductOrServicesSomeInstancesPlaceholder --> gr:SomeItems gr:LocationOfSalesOrServiceProvisioning --> gr:Location For new data and applications, please use the new, shorter elements. Old data does not have to be updated, as long as the consuming client or triplestore can handle owl:equivalentClass. If you touch your modeling patterns or queries, it would be better to move to the new identifiers now. Also, we added a few new elements: * Additional payment options: gr:JCB, gr:PaySwarm, and gr:GoogleCheckout * gr:GTIN-8 property for GTIN-8 (EAN/UCC-8) codes * gr:displayPosition (owl:AnnotationProperty) so that a preferred display position for elements and data can be represented * gr:Brand class and gr:hasBrand property for relating gr:ProductOrService and gr:BusinessEntity nodes to a brand name or brand identity For the full Changelog, see http://wiki.goodrelations-vocabulary.org/Changelog/20110401 A huge thanks to my team members Andreas Radinger, Alex Stolz, Uwe Stoll, and L?szl? T?r?k for their hard work on this! Best wishes Martin Hepp -------------- next part -------------- A non-text attachment was scrubbed... Name: Screen shot 2011-06-09 at 12.30.30 AM.png Type: image/png Size: 117336 bytes Desc: not available URL: -------------- next part -------------- From martin.hepp at ebusiness-unibw.org Thu Jun 9 22:23:41 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Thu, 9 Jun 2011 22:23:41 +0200 Subject: [goodrelations] W3C plans to add the GoodRelations to the RDFa 1.1 default profile Message-ID: Dear all: Good news for GoodRelations: The W3C plans to add the GoodRelations prefix "gr:" to the (pretty restrictive) list of predefined prefixes in the RDFa 1.1 default profile. This means that a RDFa 1.1 processor will understand GoodRelations markup (and other popular vocabularies) without the need to define the respective prefixes. This simplifies the respective markup a lot. For GoodRelations scenarios, the following are the most important prefixes included in the default profile: dcterms http://purl.org/dc/terms/ foaf http://xmlns.com/foaf/0.1/ gr http://purl.org/goodrelations/v1# rev http://purl.org/stuff/rev# v http://rdf.data-vocabulary.org/# vcard http://www.w3.org/2006/vc
Hepp Personal SCSI Controller Card Product image Average rating: 4.9, based on 99 reviews
can be simplified to
Hepp Personal SCSI Controller Card Product image Average rating: 4.9, based on 99 reviews
That is great news! Best wishes Martin Hepp Links: [1] http://www.w3.org/profile/rdfa-1.1#crawl [2] http://www.w3.org/blog/SW/2011/06/09/what_are_the_most_widely_used_vocabulari From Michael.Shepherd at xerox.com Tue Jun 14 22:26:16 2011 From: Michael.Shepherd at xerox.com (Shepherd, Michael) Date: Tue, 14 Jun 2011 13:26:16 -0700 Subject: [goodrelations] GoodRelations for Coupons Message-ID: <0131AC21C657A3418A5B6CA168C4C57E0B195031@USA7061MS04.na.xerox.net> Hi all, I was curious to know if there are current activities or potential user scenarios that use GoodRelations in the context of coupons? Is there a coupon ontology extension or perhaps a way that GoodRelations can help to specify coupons? Thanks! Mike -------------- next part -------------- An HTML attachment was scrubbed... URL: From ltorokjr at gmail.com Wed Jun 15 11:18:03 2011 From: ltorokjr at gmail.com (=?UTF-8?B?TMOhc3psw7MgVMO2csO2aw==?=) Date: Wed, 15 Jun 2011 11:18:03 +0200 Subject: [goodrelations] GoodRelations for Coupons In-Reply-To: <0131AC21C657A3418A5B6CA168C4C57E0B195031@USA7061MS04.na.xerox.net> References: <0131AC21C657A3418A5B6CA168C4C57E0B195031@USA7061MS04.na.xerox.net> Message-ID: Hi Michael, I think it really depends what aspect of coupons you want to model. In the context of selling the coupons (a la Groupon), I think coupons are just like any other product. So you could create a gr:Offering with a proper price specification and the coupon as product attached to it. The product details (e.g. how to redeem) can be attached as text (gr:Description). Coupons are (usually) also a discount buy option that you can be exercised later. What I mean by that is they seem to me as a materialization of a gr:Offering. I don't think you can currently say explicitly that an instance of gr:Offering can be bought or sold. Coupons can also be distinct. Although instances of gr:Offering are usually modeled as blank nodes, as countability is not a concern. (You model it implicitly through gr:hasInventoryLevel), however, if I'm interested in identifying and track a coupon I'd give it URI. All in all, GR can be used to model coupons as products to be sold. I'm not sure about the second part of my thought exercise though. Bests, Laszlo 2011/6/14 Shepherd, Michael > Hi all, > > > > I was curious to know if there are current activities or potential user > scenarios that use GoodRelations in the context of coupons? Is there a > coupon ontology extension or perhaps a way that GoodRelations can help to > specify coupons? > > > > Thanks! > > Mike > > > > _______________________________________________ > goodrelations mailing list > goodrelations at ebusiness-unibw.org > http://ebusiness-unibw.org/cgi-bin/mailman/listinfo/goodrelations > > -- L?szl? T?r?k Skype: laczoka2000 Twitter: @laczoka -------------- next part -------------- An HTML attachment was scrubbed... URL: From martin.hepp at ebusiness-unibw.org Wed Jun 15 12:13:17 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Wed, 15 Jun 2011 12:13:17 +0200 Subject: [goodrelations] GoodRelations for Coupons In-Reply-To: <0131AC21C657A3418A5B6CA168C4C57E0B195031@USA7061MS04.na.xerox.net> References: <0131AC21C657A3418A5B6CA168C4C57E0B195031@USA7061MS04.na.xerox.net> Message-ID: Hi Michael, Thanks for your question. There are several ways of modeling coupon-information in GoodRelations: > > Hi all, > > I was curious to know if there are current activities or potential user scenarios that use GoodRelations in the context of coupons? Is there a coupon ontology extension or perhaps a way that GoodRelations can help to specify coupons? > 1. Gift Vouchers If the coupon has a money equivalent and is basically a method of payment, then your type of coupon is an instance of gr:PaymentMethod. Attach this type to the offers for which it can be turned in. foo:GiftVoucher a gr:PaymentMethod ; gr:name "ACME Gift Vouchers (gr:PaymentMethod)" . foo:Offer0001 a gr:Offering ; gr:acceptedPaymentMethods foo:GiftVoucher. You may then also want to model that one can buy gift vouchers, which would happen by creating a product class foo:ACMEGiftVouchers a owl:Class ; rdfs:subClassOf gr:ProductOrService ; gr:name "ACME Gift Vouchers" . and two properties foo:value a owl:DatatypeProperty ; rdfs:subPropertyOf gr:datatypeProductOrServiceProperty . foo:currency a owl:DatatypeProperty ; rdfs:subPropertyOf gr:datatypeProductOrServiceProperty . Then, you can offer individual gift vouchers for sale: foo:GiftVoucher20USD a foo:ACMEGiftVouchers, gr:SomeItems ; foo:value 20.00 ; foo:currency "USD"^^xsd:string . foo:Offer0002 a gr:Offering ; gr:includes foo:GiftVoucher20USD ; gr:hasPriceSpecification [a gr:UnitPriceSpecification ; gr:hasCurrency "USD"^^xsd:string ; gr:hasCurrencyValue 20.00 ; gr:validThrough "2011-31-12"^^xsd:datetime ] . 2. Coupons that entitle you to obtain a product or service at a reduced price (in the Groupon sense) In this case, you could model a special offer (gr:Offering) and constrain it's validity to a special audience via gr:eligibleCustomerTypes. foo:SummerSaleCoupon a gr:BusinessEntityType ; gr:name "Summer Sale Coupon must be presented" . foo:Offer0003 a gr:Offering ; gr:eligibleCustomerTypes foo:SummerSaleCoupon ; gr:hasPriceSpecification [a gr:UnitPriceSpecification ; gr:hasCurrency "USD"^^xsd:string ; gr:hasCurrencyValue 10.00 ; gr:validThrough "2011-31-12"^^xsd:datetime ] . I have been thinking about a special coupon extension for GoodRelations; this is on our radar. If there is significant interest, i.e. potential real-world adopters, we can make this a priority. Best wishes Martin Hepp On Jun 14, 2011, at 10:26 PM, Shepherd, Michael wrote: > Hi all, > > I was curious to know if there are current activities or potential user scenarios that use GoodRelations in the context of coupons? Is there a coupon ontology extension or perhaps a way that GoodRelations can help to specify coupons? > > Thanks! > Mike > > _______________________________________________ > goodrelations mailing list > goodrelations at ebusiness-unibw.org > http://ebusiness-unibw.org/cgi-bin/mailman/listinfo/goodrelations From s.abdel-halim at student.HTW-Berlin.de Wed Jun 15 12:13:11 2011 From: s.abdel-halim at student.HTW-Berlin.de (Sharif Abdel-Halim) Date: Wed, 15 Jun 2011 12:13:11 +0200 Subject: [goodrelations] Fragen zu GoodRelations Message-ID: Guten Tag, erlauben Sie mir bitte kurz mich vorzustellen: Ich hei?e Sharif Abdel-Halim und bin Student der angewandten Informatik auf der Hochschule f?r Technik und Wirtschaft. Im Rahmen einer Studienarbeit soll ich eine Ontologie f?r Produkte der Automatisierungstechnik entwerfen. Ich bin zuf?llig auf das GoodRelations Vokabular gesto?en und finde die gesamte Vision, die dahinter steckt, sehr interessant. Ich habe mir bereits einiges dazu durchgelesen und angeh?rt. Ich habe auch schon diverse Tools gesehen, mit denen man beispielsweise eine Umformung von XML-Dateien in RDFa durchf?hren kann. Das Problem dabei ist nur, dass sowas in meinem Fall bisher nicht vorliegt. Das hei?t, dass bis jetzt noch keine strukturieren Daten vorliegen. Ich muss sie zu erst selbst strukturieren. Dabei stellt sich die Frage, ob ich das mit einem bekannten Framework wie dem Jena Framework machen soll oder auf ein solches Framework verzichten soll. Die Frage ist wobei hier die Vor-und Nachteile bei der Verwendung solcher Frameworks liegen. Eines der Vorteile ist sicherlich, dass sie bereits ?ber eine XML-Schnittstelle verf?gen und ?ber Inferenzmechanimen, die ich nicht selbst implementieren m?sste. Mir ist zu Augen gekommen, dass sie hierf?r eine API zur Verf?gung stellen oder zumindest gerade entwickeln: Die Jenagr um GoodRelations in Jena integrieren zu k?nnen. Leider habe ich keine n?here Informationen dazu finden k?nnen. Auch stehen noch keine Sourcefiles zum Download zur Verf?gung. Ist das alles noch in Arbeit oder gibt es bereits erste Versionen davon?? Des Weiteren hatte ich eine Frage zu der GoodRelations Erweiterung f?r den Firefox Browser. Ich habe es installiert aber wenn ich etwas markiere und dann auf Firmensuche oder Produktsuche, auf einem auf RDFa basierenden Produktkatalog wie bei Jarltech, gehe, passiert leider gar nichts. Gibt es vielleicht dazu n?here Informationen dazu wie man es richtig verwendet? Ich w?re ?ber eine Antwort und eine Hilfe sehr dankbar. Mit freundlichen Gr??en Sharif Abdel-Halim From martin.hepp at ebusiness-unibw.org Wed Jun 15 14:24:38 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Wed, 15 Jun 2011 14:24:38 +0200 Subject: [goodrelations] Fragen zu GoodRelations In-Reply-To: References: Message-ID: <7F9077A2-4A2F-4112-B151-472B02BF0F59@ebusiness-unibw.org> Dear Mr. Abdel-Halim, unfortunately, the GoodRelations mailing list is in English only. If you want to build your own GoodRelations-compliant ontology for product or service *types* and *properties*, this is the recipe you need: http://www.ebusiness-unibw.org/wiki/Own_GoodRelations_Vocabularies I am not aware of a Java/Jena-Framework for producing or consuming GoodRelations data, but there is one for PHP: http://code.google.com/p/gr4php/ But basically, you can use plain Jena for creating GoodRelations data. The process of exposing rich product model master data ("datasheets") using GoodRelations is described at http://www.ebusiness-unibw.org/wiki/Modeling_Product_Models The Firefox extension is currently accepting GoodRelations in RDFa only. Jarltech uses GoodRelations in RDF/XML syntax. Best wishes Martin Hepp On Jun 15, 2011, at 12:13 PM, Sharif Abdel-Halim wrote: > Guten Tag, > erlauben Sie mir bitte kurz mich vorzustellen: Ich hei?e Sharif > Abdel-Halim und bin Student der angewandten Informatik auf der Hochschule > f?r Technik und Wirtschaft. Im Rahmen einer Studienarbeit soll ich eine > Ontologie f?r Produkte der Automatisierungstechnik entwerfen. > Ich bin zuf?llig auf das GoodRelations Vokabular gesto?en und finde die > gesamte Vision, die dahinter steckt, sehr interessant. Ich habe mir > bereits einiges dazu durchgelesen und angeh?rt. > Ich habe auch schon diverse Tools gesehen, mit denen man beispielsweise > eine Umformung von XML-Dateien in RDFa durchf?hren kann. > Das Problem dabei ist nur, dass sowas in meinem Fall bisher nicht > vorliegt. Das hei?t, dass bis jetzt noch keine strukturieren Daten > vorliegen. Ich muss sie zu erst selbst strukturieren. Dabei stellt sich > die Frage, ob ich das mit einem bekannten Framework wie dem Jena Framework > machen soll oder auf ein solches Framework verzichten soll. Die Frage ist > wobei hier die Vor-und Nachteile bei der Verwendung solcher Frameworks > liegen. Eines der Vorteile ist sicherlich, dass sie bereits ?ber eine > XML-Schnittstelle verf?gen und ?ber Inferenzmechanimen, die ich nicht > selbst implementieren m?sste. > Mir ist zu Augen gekommen, dass sie hierf?r eine API zur Verf?gung stellen > oder zumindest gerade entwickeln: Die Jenagr um GoodRelations in Jena > integrieren zu k?nnen. Leider habe ich keine n?here Informationen dazu > finden k?nnen. Auch stehen noch keine Sourcefiles zum Download zur > Verf?gung. Ist das alles noch in Arbeit oder gibt es bereits erste > Versionen davon?? > Des Weiteren hatte ich eine Frage zu der GoodRelations Erweiterung f?r den > Firefox Browser. Ich habe es installiert aber wenn ich etwas markiere und > dann auf Firmensuche oder Produktsuche, auf einem auf RDFa basierenden > Produktkatalog wie bei Jarltech, gehe, passiert leider gar nichts. > Gibt es vielleicht dazu n?here Informationen dazu wie man es richtig > verwendet? > Ich w?re ?ber eine Antwort und eine Hilfe sehr dankbar. > > Mit freundlichen Gr??en > > Sharif Abdel-Halim > > > > > _______________________________________________ > goodrelations mailing list > goodrelations at ebusiness-unibw.org > http://ebusiness-unibw.org/cgi-bin/mailman/listinfo/goodrelations From martin.hepp at ebusiness-unibw.org Wed Jun 15 21:44:49 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Wed, 15 Jun 2011 21:44:49 +0200 Subject: [goodrelations] Google bug: foaf:page and Google Rich Snippets Message-ID: <4C3E2A83-037E-4A13-8A3C-328F346A16DA@ebusiness-unibw.org> Hi all: I just found out that the Google Rich Snippets Testing tool rejects all RDFa markup for products if the data contains foaf:page (a link to the HTML page) with a value that is NOT from within the domain name space of the actual file location. Example: If you have the following markup in the page available from http://www.domain.com/example.html, then a) an empty foaf:page value will work
b) foaf:page with the actual page URI will work
c) foaf:page with a URI from the same domain name space as the actual page URI will also work:
BUT: c) foaf:page with a URI from another domain name space than the actual page URI will NOT work:
Unfortunately, the testing tool gives the very uninformative error message "Insufficient data to generate the preview. ... Warning: In order to generate a preview, either price or review or availability needs to be present." It is natural that Google does not like foaf:page pointers to external domain name spaces, but this unexpected behavior costed me almost a day. So watch out! Option b) is the recommended practice, because it works in any scenario, but it requires that you know the URI of the final publishing destination at markup-time (or the template has access to the deep URI as a variable). Option a) is handier, because the URI of the current page will be automatically used as the value for foaf:page, but it fails, if the document sets the base URI of all product item pages to the main shop page, e.g. via because then all offers would point to the same base page and deep links would be lost. Best wishes Martin Hepp From beroca at gmail.com Mon Jun 20 12:55:12 2011 From: beroca at gmail.com (Bene Rodriguez-Castro) Date: Mon, 20 Jun 2011 11:55:12 +0100 Subject: [goodrelations] Fwd: Answers - SemanticWeb: New question: How does one publish semantic data for example GoodRelations? on ANSWERS In-Reply-To: <201106200035.p5K0ZmQ4002602@wmb2.iworld.com> References: <201106200035.p5K0ZmQ4002602@wmb2.iworld.com> Message-ID: Interesting question from answers.semanticweb.com, in case some GR users may want to share their experiences or some tips. ----- How does one publish semantic data for example GoodRelations? http://answers.semanticweb.com/questions/10286/how-does-one-publish-semantic-data-for-example-goodrelations ----- Best, Bene Rodriguez School of Electronics and Computer Science University of Southampton Southampton, SO17 1BJ, United Kingdom ---------- Forwarded message ---------- From: ANSWERS Date: Mon, Jun 20, 2011 at 1:35 AM Subject: Answers - SemanticWeb: New question: How does one publish semantic data for example GoodRelations? on ANSWERS To: beroca [image: Answers.SemanticWeb.com] ------------------------------ Hello beroca, hatescheese has just posted a new question on ANSWERS, entitled How does one publish semantic data for example GoodRelations?and tagged " *rdfa goodrelations *". Here's what it says: Hi all, I am sorry if this seems like a dumb question, but I am interested in using GoodRelations offers with rdfa (Very New to Both) but i can't find any documentation on their site. It talks about how it opens up your data to various channels such as mobile applications, and browser extensions for example, but it doesn't say how you get your data on there. Once you mark up your pages do you have to submit a sitemap somewhere or submit a url somewhere? Thanks in advance Don't forget to come over and cast your vote. Thanks, ANSWERS P.S. You can always fine-tune which notifications you receive here. ------------------------------ MediaBistro | 475 Park Avenue South, 4th Floor | New York, New York 10016 | Phone: 212-389-2000 -------------- next part -------------- An HTML attachment was scrubbed... URL: From kredfield at juniper.net Mon Jun 20 20:10:56 2011 From: kredfield at juniper.net (Keith Redfield) Date: Mon, 20 Jun 2011 11:10:56 -0700 Subject: [goodrelations] Semantic markup for knowledge/information? Message-ID: <09C251EE7F952E4BA09569EBE30B2349414DD301AA@EMBX01-HQ.jnpr.net> Hi, This may be a bit OT, since GoodRelations seems to have it's feet firmly in ecommerce. But I am wondering if anyone on-list has insight into how semantic markup might be used/levered in knowledge base(KB) / knowledge-management domains, such as enterprise technical support. Clearly we could add product markup to KB articles _about_ a product. And that's somewhat helpful all by itself. But concepts other than product are (usually) contained within - for example in our case, technologies, features, specifications, limitations, etc. Then my other question is - if we had something, so what? It's not clear to me that the recent advances with Google et al would immediately benefit. But I imagine that if all high-tech vendors (for example) whose products have to work together (Microsoft, Apple, IBM, Juniper, HP, Oracle etc) adopted such markup - it might eventually be easier for consumers (in the literal sense) to find answers. (A simple example would be BestBuy being able to reference all articles at apple.com relevant to the iPad you're browsing...but also even the articles at Juniper.net describing our iPad app) I have only a lay-persons interest and knowledge of the topic, so apologies if this is way off base, but any re-directs appreciated. Regards, Keith Redfield Juniper Networks, Inc. And btw we will hopefully soon have some GR markup on our product support pages @ www.juniper.net -------------- next part -------------- An HTML attachment was scrubbed... URL: From semantics at 0x1b.com Mon Jun 20 21:21:52 2011 From: semantics at 0x1b.com (Ed - 0x1b, Inc.) Date: Mon, 20 Jun 2011 12:21:52 -0700 Subject: [goodrelations] Semantic markup for knowledge/information? In-Reply-To: <09C251EE7F952E4BA09569EBE30B2349414DD301AA@EMBX01-HQ.jnpr.net> References: <09C251EE7F952E4BA09569EBE30B2349414DD301AA@EMBX01-HQ.jnpr.net> Message-ID: Maybe something like SIOC would be great long term, for simple/now - you could use (Open)Calais for tagging KB articles - might be a start. or http://semanticweb.com/wand-releases-new-it-department-taxonomy_b20713 if you can't frame it as a product offer or request then GoodRelations is likely OT, not that you can't have GoodRelations RDF in with your other RDF - open world and all that. so, how's that IPv6 thing coming .... ;) Ed On Mon, Jun 20, 2011 at 11:10 AM, Keith Redfield wrote: > Hi, > > This may be a bit OT, since GoodRelations seems to have it?s feet firmly in > ecommerce. But I am wondering if anyone on-list has insight into how > semantic markup might be used/levered in knowledge base(KB) / > knowledge-management domains, such as enterprise technical support. Clearly > we could add product markup to KB articles _about_ a product. And that?s > somewhat helpful all by itself. But concepts other than product are > (usually) contained within ? for example in our case, technologies, > features, specifications, limitations, etc. > > Then my other question is ? if we had something, so what? It?s not clear to > me that the recent advances with Google et al would immediately benefit. But > I imagine that if all high-tech vendors (for example) whose products have to > work together (Microsoft, Apple, IBM, Juniper, HP, Oracle etc) adopted such > markup ? it might eventually be easier for consumers? (in the literal sense) > to find answers. (A simple example would be BestBuy being able to reference > all articles at apple.com relevant to the iPad you?re browsing?but also even > the articles at Juniper.net describing our iPad? app) > > I have only a lay-persons interest and knowledge of the topic, so apologies > if this is way off base, but any re-directs appreciated. > > Regards, > > Keith Redfield > Juniper Networks, Inc. > > And btw we will hopefully soon have some GR markup on our product support > pages @ www.juniper.net > > _______________________________________________ > goodrelations mailing list > goodrelations at ebusiness-unibw.org > http://ebusiness-unibw.org/cgi-bin/mailman/listinfo/goodrelations > > From martin.hepp at ebusiness-unibw.org Tue Jun 21 09:07:27 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Tue, 21 Jun 2011 09:07:27 +0200 Subject: [goodrelations] Google / prices in Euro - Bug & Work-around Message-ID: <884B7204-9A97-4E6B-B735-A1A450BF443C@ebusiness-unibw.org> Dear all: There is a small bug in Google's Rich Snippet testing tool: If the currency is "EUR", then the price may be rendered incorrectly - in some cases, the digits following the delimiter (99.xx - the fractional part) are treated as if they were part of the main digits (i.e. the integer part). Example: foo:price a gr:UnitPriceSpecification ; gr:hasCurrency "EUR"^^xsd:string ; gr:hasCurrencyValue "99.00"^^xsd:float . can, in some cases, show up as EUR 9900.00 Google is working on the issue and we assume at this point that this is just a bug in the validator, not in the actual Google rendering component. Other currencies are not affected, as far as we know. As a work-around, we recommend that you remove the fractional part from currency values if it is ".00". So for those values, please use foo:price a gr:UnitPriceSpecification ; gr:hasCurrency "EUR"^^xsd:string ; gr:hasCurrencyValue "99.00"^^xsd:float . or in RDFa
instead of
This is just a precautionary measure. At this moment, we have not seen this bug in actual Google search results, just in the testing tool. Best wishes Martin Hepp From martin.hepp at ebusiness-unibw.org Thu Jun 23 11:14:39 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Thu, 23 Jun 2011 11:14:39 +0200 Subject: [goodrelations] Semantic markup for knowledge/information? In-Reply-To: <09C251EE7F952E4BA09569EBE30B2349414DD301AA@EMBX01-HQ.jnpr.net> References: <09C251EE7F952E4BA09569EBE30B2349414DD301AA@EMBX01-HQ.jnpr.net> Message-ID: <39882DB8-F97D-4976-B074-92743F931E21@ebusiness-unibw.org> Hi Keith, all: What I would recommend is: 1. Create a URI for a product model, e.g. foo:iPhone3GS-8GB a gr:ProductOrServiceModel ; .... features come in here ... ; foaf:page . 2. Then, a document related to this product can be linked to that make and model via foaf:topic or foaf:primaryTopic: foo:IPhoneFAQ a a foaf:Document; foaf:primaryTopic foo:iPhone3GS-8GB. Best Martin On Jun 20, 2011, at 8:10 PM, Keith Redfield wrote: > Hi, > > This may be a bit OT, since GoodRelations seems to have it?s feet firmly in ecommerce. But I am wondering if anyone on-list has insight into how semantic markup might be used/levered in knowledge base(KB) / knowledge-management domains, such as enterprise technical support. Clearly we could add product markup to KB articles _about_ a product. And that?s somewhat helpful all by itself. But concepts other than product are (usually) contained within ? for example in our case, technologies, features, specifications, limitations, etc. > > Then my other question is ? if we had something, so what? It?s not clear to me that the recent advances with Google et al would immediately benefit. But I imagine that if all high-tech vendors (for example) whose products have to work together (Microsoft, Apple, IBM, Juniper, HP, Oracle etc) adopted such markup ? it might eventually be easier for consumers (in the literal sense) to find answers. (A simple example would be BestBuy being able to reference all articles at apple.com relevant to the iPad you?re browsing?but also even the articles at Juniper.net describing our iPad app) > > I have only a lay-persons interest and knowledge of the topic, so apologies if this is way off base, but any re-directs appreciated. > > Regards, > > Keith Redfield > Juniper Networks, Inc. > > And btw we will hopefully soon have some GR markup on our product support pages @ www.juniper.net > > _______________________________________________ > goodrelations mailing list > goodrelations at ebusiness-unibw.org > http://ebusiness-unibw.org/cgi-bin/mailman/listinfo/goodrelations From martin.hepp at ebusiness-unibw.org Thu Jun 23 15:22:35 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Thu, 23 Jun 2011 15:22:35 +0200 Subject: [goodrelations] Reviews / Bug in Google's documentation References: <30CA0A3E-AD14-47DA-B131-E5C46F0C56BA@gmail.com> Message-ID: <2321DCB8-AE90-4C4C-8579-AC8E98B54D26@ebusiness-unibw.org> Dear all: I just spotted a severe bug in the Google recipe for reviews, given at [1]. Please fix this in your implementations asap, because otherwise, the snippet in the Google search results may show less stars than it should. Instead of v:rating for the average, use v:average. If you use v:rating for aggregate reviews, Google seems to divide it by the number of reviews, so you may get only one star even if your average is 4.4. The following would be a correct review pattern that works with both Google and Yahoo in RDFa:
I have already asked Google to fix this in their documentation. So please fix it in your code asap. Thanks! Best wishes Martin Hepp [1] http://www.google.com/support/webmasters/bin/answer.py?answer=186036 From martin.hepp at ebusiness-unibw.org Thu Jun 23 15:40:37 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Thu, 23 Jun 2011 15:40:37 +0200 Subject: [goodrelations] and relative URIs Message-ID: Dear all: I just found out that some sites use the element in shop pages for defining the base URI for relative references (e.g. "#offer"). Be careful when using in combination with RDFa, because you may spoil the identifiers of your data objects: For instance, means that Payment Options will link to http://www.myshop.com/info/payment This is often the intended effect. Unfortunately, it will also mean that an RDFa element with a relative identifier, e.g.
... will get the global URI http://www.myshop.com/info#offer Now, since you usually have many different offers they will all get the same URI, which renders your data useless. So if you want to use with RDFa, then you must choose one of the following approaches: 1. Take the true ("canonical") URI of the individual page as the base URI, e.g. OR 2. Use full URIs for all "about" properties, e.g.
If you send the "deep" URI of the actual page to the template engine as a parameter anyway, you could also write
or similar. The exact syntax will depend on your templating engine, of course. Best wishes Martin Hepp From martin.hepp at ebusiness-unibw.org Fri Jun 24 08:36:12 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Fri, 24 Jun 2011 08:36:12 +0200 Subject: [goodrelations] SEOmoz post on GoodRelations Message-ID: <39BFFE5D-CA88-4652-9872-78C52CF682EC@ebusiness-unibw.org> Dear all: I just posted additional information on the relationship between GoodRelations and schema.org on SEOmoz. This may be of general interest, so I repost it here: -----snip ----- I think it is utterly needed to clarify the relationship between - GoodRelations vs. schema.org elements - Microdata vs. RDFa 1. Google and Yahoo have confirmed explicitly that GoodRelations in RDFa syntax remains a fully supported markup for product / e-commerce data. Bing has also just announced they will support GoodRelations in the future. 2. Since mixing RDFa and Microdata is not recommended by Google and Yahoo, GoodRelations in RDFa is the only proper way of combining Facebook Open Graph markup with markup for rich snippets, since Facebook is not part of the party. 3. The Rich Snippets Validator (example: http://www.google.com/webmasters/tools/richsnippets?url=http%3A%2F%2Fwww.heppnetz.de%2Frdfa4google%2Ftc1.html&view=) can currently show a preview only for GoodRelations in RDFa, not for schema.org in Microdata. That means that you can currently debug your markup only for GoodRelations in RDFa; for Microdata, you have no effective testing environment. 4. Google extracts way more GoodRelations properties than listed in the short documentation on the Google site. Check here for extended testcases: http://www.heppnetz.de/rdfa4google/testcases.html They all validate with Google and Yahoo. Now, the interesting thing is that you can bet that Google uses this additional information as a signal for future relevance assessment. For instance, if you explicitly state that your offer is valid in California using gr:eligibleRegions, e.g.
We deliver to California!
...
then you can bet that a request from an IP in California is more likely to see this than an offer that lacks target region information. 5. Even if you want to use the main schema.org elements in Microdata syntax, you can combine them with GoodRelations properties and classes for things that schema.org does not support or supports only poorly, like - shipping costs, - payment options, - structured opening hours, etc. We will add respective Microdata examples to the GoodRelations documentation at http://purl.org/goodrelations/v1.html shortly. Best wishes Martin Hepp From martin.hepp at ebusiness-unibw.org Fri Jun 24 09:22:04 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Fri, 24 Jun 2011 09:22:04 +0200 Subject: [goodrelations] Reviews / Bug in Google's documentation In-Reply-To: <2321DCB8-AE90-4C4C-8579-AC8E98B54D26@ebusiness-unibw.org> References: <30CA0A3E-AD14-47DA-B131-E5C46F0C56BA@gmail.com> <2321DCB8-AE90-4C4C-8579-AC8E98B54D26@ebusiness-unibw.org> Message-ID: <5BB1F1DA-A36B-401B-B19C-82464BDEB861@ebusiness-unibw.org> Dear all: Apologies for the confusion - the main cause of the problem is that Google is also accepting properties from the Yahoo review vocabulary http://purl.org/stuff/rev# without telling us ;-) And the min/max rating in that vocabulary is underspecified - I thought it indicates the range of existing ratings (e.g. from 3 to 5), but Google seems to use as the range of the possible ratings. The unfortunate effect is that a 4.5 rating with min=4 and max=5 means that Google would display only 2.5 stars on a 5-star-scale, because 4.5 is in the middle between 4 and 5. So the official recommendation from our side is now to NOT USE review:minRating and review:maxRating. Simply delete the two lines: >
>
The complete pattern is > >
xmlns:v="http://rdf.data-vocabulary.org/#" > typeof="gr:Offering" about="#offer"> >
>
>
>
>
>
>
Still, v:average seems to be the correct property instead of v:rating for Google. But we asked them for clarification. Our tests show that both seem to work. Best Martin Hepp On Jun 23, 2011, at 3:22 PM, Martin Hepp wrote: > Dear all: > > I just spotted a severe bug in the Google recipe for reviews, given at [1]. Please fix this in your implementations asap, because otherwise, the snippet in the Google search results may show less stars than it should. > > Instead of v:rating for the average, use v:average. If you use v:rating for aggregate reviews, Google seems to divide it by the number of reviews, so you may get only one star even if your average is 4.4. > > The following would be a correct review pattern that works with both Google and Yahoo in RDFa: > >
xmlns:v="http://rdf.data-vocabulary.org/#" > typeof="gr:Offering" about="#offer"> >
>
>
>
>
>
>
>
>
> > I have already asked Google to fix this in their documentation. > > So please fix it in your code asap. Thanks! > > Best wishes > > Martin Hepp > > [1] http://www.google.com/support/webmasters/bin/answer.py?answer=186036 > > > > > _______________________________________________ > goodrelations mailing list > goodrelations at ebusiness-unibw.org > http://ebusiness-unibw.org/cgi-bin/mailman/listinfo/goodrelations From martin.hepp at ebusiness-unibw.org Fri Jun 24 10:47:39 2011 From: martin.hepp at ebusiness-unibw.org (Martin Hepp) Date: Fri, 24 Jun 2011 10:47:39 +0200 Subject: [goodrelations] Bandwidth management for large RDF resources Message-ID: Dear all: When you publish large amounts of GoodRelations in dump files, e.g. in RDF/XML or Ntriples syntax, you should take measures against excessive traffic caused by crawlers and other clients. The first set of measures is to properly support caching and avoid unnecessary requests. In particular, you should 1. properly set the lastmod attribute in your sitemap.xml, i.e., avoid using the date of creating the sitemap for all entries 2. properly configure caching information sent in the HTTP response header. Second, if you host really significant amounts of data, you should limit the maximum download speed for certain resources and exclude bad bots. Here is a good resource on this for Apache environments: http://www.whoopis.com/howtos/web-bandwidth-limit.html#bwmod For instance, you could throttle the download speed for large .rdf files to 10 KB/ sec: # RDF/XML files larger than 1MB go at 10k/sec max LargeFileLimit .rdf 1000 10000 Best wishes Martin Hepp