On Wed, Dec 10, 2014 at 4:32 PM, Dimitris Kontokostas <[email protected]> wrote: > > > On Wed, Dec 10, 2014 at 5:07 PM, Magnus Knuth > <[email protected]> wrote: >> >> Just an add-on about what I did not understand. >> >> Asking for the resource of Park Güell in DBpedia I have to use the >> URI-encoded URL: >> curl -L -H "Accept: text/turtle; q=1.0, application/x-turtle; q=0.9, >> text/n3; q=0.8, application/rdf+xml; q=0.5, text/plain; q=0.1" >> http://dbpedia.org/resource/Park_G%C3%BCell >> >> The IRI-encoded URL does not return any result: >> curl -L -H "Accept: text/turtle; q=1.0, application/x-turtle; q=0.9, >> text/n3; q=0.8, application/rdf+xml; q=0.5, text/plain; q=0.1" >> http://dbpedia.org/resource/Park_Güell
HTTP only allows ASCII chars in the request path. It's unclear what curl is doing with the non-ASCII character "ü", and it's unspecified what the server is doing with the non-ASCII bytes that curl is presumably sending. For valid HTTP requests, IRIs must be converted to URIs. This is an area where RDF and HTTP are at odds. >> >> Why that, it should at least return the owl:sameAs >> dbpedia:Park_G%C3%BCell, right? >> >> Though for German DBpedia both URLs (IRI and URI) return exactly the same: >> curl -L -H "Accept: text/turtle; q=1.0, application/x-turtle; q=0.9, >> text/n3; q=0.8, application/rdf+xml; q=0.5, text/plain; q=0.1" >> http://de.dbpedia.org/resource/Park_Güell >> curl -L -H "Accept: text/turtle; q=1.0, application/x-turtle; q=0.9, >> text/n3; q=0.8, application/rdf+xml; q=0.5, text/plain; q=0.1" >> http://de.dbpedia.org/resource/Park_G%C3%BCell Apparently, your instance of curl and the HTTP server at de.dbpedia.org happen to agree on the encoding of non-ASCII characters in the request line. In other words, they both happen to break the HTTP spec in the same way. :-) > > > The reason for this behavior is the 'dbp_decode_iri' registry value. See the > following links for examples > https://github.com/dbpedia/dbpedia-vad-i18n/blob/master/dbpedia/dbpedia_local.sql#L81 > https://github.com/dbpedia/dbpedia-vad-i18n/blob/master/dbpedia/vsp/description.vsp#L215 > https://github.com/dbpedia/dbpedia-vad-i18n/blob/master/dbpedia/dbpedia_local.sql#L314 > > In English this variable is set to false and thus, IRIs are not decoded > properly to get resolved. The idea is that HTTP protocol encodes all > requests to URIs and then the server (vos) is responsible for resolving to > the correct resource to a URI (leaving the request as is) or to an IRI (by > decoding the request). The trick here is that when we enable IRIs we cannot > know if a request was about an IRI or a URI since HTTP converts everything > to URIs, thus both links will work > >> >> Thing is that the German DBpedia points to owl:sameAs dbpedia:Park_Güell >> and most other language chapters do that. >> >> I totally understand the problem of breaking legacy links from external >> datasets. But we should somehow discuss how a transformation can be >> achieved, because as said the problem gets not solved by postponing it year >> after year. We also should consider that quite a bunch of external datasets >> are already broken, due to changes in the URI structures (e.g. usage of >> unescaped comma and brackets) or resources that disappeared over time. >> One solutions could be redirects for dereferencing. But also accessing >> DBpedia via SPARQL should be considered. > > > I also agree that leaving this doesn't solve the problem. I think we should > make the switch in the next release but let's think of the best way to do > it. > >> >> Finally, as I understood now (it is not mentioned on >> [http://wiki.dbpedia.org/DatasetsLoaded2014] properly) the .nt (URI) and not >> the .ttl (IRI) datasets have been imported!? >> >> If you like, we could start a group discussion, collecting all possible >> issues and effects related to that, and looking for viable solutions. >> >> Thanks, >> Magnus >> >> -- >> Magnus Knuth >> >> Hasso-Plattner-Institut für Softwaresystemtechnik GmbH >> Prof.-Dr.-Helmert-Str. 2-3 >> 14482 Potsdam >> >> Amtsgericht Potsdam, HRB 12184 >> Geschäftsführung: Prof. Dr. Christoph Meinel >> >> tel: +49 331 5509 547 >> email: [email protected] >> web: http://www.hpi.de/ >> webID: http://magnus.13mm.de/ >> >> >> >> ------------------------------------------------------------------------------ >> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server >> from Actuate! Instantly Supercharge Your Business Reports and Dashboards >> with Interactivity, Sharing, Native Excel Exports, App Integration & more >> Get technology previously reserved for billion-dollar corporations, FREE >> >> http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk >> _______________________________________________ >> Dbpedia-developers mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/dbpedia-developers > > > > > -- > Dimitris Kontokostas > Department of Computer Science, University of Leipzig > Research Group: http://aksw.org > Homepage:http://aksw.org/DimitrisKontokostas > > ------------------------------------------------------------------------------ > Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server > from Actuate! Instantly Supercharge Your Business Reports and Dashboards > with Interactivity, Sharing, Native Excel Exports, App Integration & more > Get technology previously reserved for billion-dollar corporations, FREE > http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk > _______________________________________________ > Dbpedia-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/dbpedia-developers > ------------------------------------------------------------------------------ Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server from Actuate! Instantly Supercharge Your Business Reports and Dashboards with Interactivity, Sharing, Native Excel Exports, App Integration & more Get technology previously reserved for billion-dollar corporations, FREE http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk _______________________________________________ Dbpedia-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dbpedia-developers
