- From: Eric Prud'hommeaux <eric@w3.org>
- Date: Thu, 17 Feb 2011 18:02:04 -0500
- To: Ivan Herman <ivan@w3.org>
- Cc: Percy Enrique Rivera Salas <privera.salas@gmail.com>, RDB2RDF WG <public-rdb2rdf-wg@w3.org>
* Eric Prud'hommeaux <eric@w3.org> [2011-02-16 10:27-0500] > * Ivan Herman <ivan@w3.org> [2011-02-16 14:47+0100] > > Actually, my mail may have been just wasted bandwidth, a.k.a. noise, because the DM explicitly refers to IRI-s already (which is perfectly fine with me!) and so does R2RML. > > Well, I think we still need to decide something here. Right now, the rule is urlencode everything. This re-uses conventional tools very well, but handles safely but indelicately. > > http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases_v1#Direct_Graph_I18NnoSpecialCharsDG Expected Result 1 shows the impact of urlencoding everything. Expected Result 2 shows what happens if we only escape the characters we need to, i.e. [,#=/>] . http://www.w3.org/2001/sw/rdb2rdf/wiki/R2RML_Test_Cases_v1#I18NnoSpecialChars now has multi-col primary and foreign keys. Pasting that SQL into http://this-db-really.does-not-exist.org/?minEncode=on will give you a relatively pretty graph, while pasting it into http://this-db-really.does-not-exist.org/ will feel like sand in your teeth. > > Ivan > > > > > > On Feb 16, 2011, at 14:01 , Percy Enrique Rivera Salas wrote: > > > > > Dear Ivan, > > > > > > I guess IRI should be the best choice, following what the other groups have done > > > (OWL2, SPARQL, RIF, RDFa) let me reference a discussion about this topic > > > in the "Semantic Web Mailing List" in the following link > > > > > > http://lists.w3.org/Archives/Public/semantic-web/2010Nov/0036.html > > > > > > Best Regards, > > > > > > Percy > > > > > > 2011/2/16 Ivan Herman <ivan@w3.org> > > > Guys, > > > > > > a simple question. What happens if the column name contains non basic ASCII characters? Ie, accented Latin, Russian, Chinese...? > > > > > > For Direct Mapping: We have to decide whether we use IRIs or whether the default is to convert IRIs into URIs according to the specs. My instinct would be to generate IRIs, but I am not sure the tooling around us is good enough. > > > > > > For R2RML: I am not familiar enough with SQL to know whether its usage to generate a virtual table covers all the possibilities. Ie, if I want to convert into URI, is it a trivial thing to do? Is it worth having some shorthands like the templates? > > > > > > This may not be a serious issue. But it is certaintly worth writing it down somewhere in the documents > > > > > > Ivan > > > > > > ---- > > > Ivan Herman > > > Tel:+31 641044153 > > > http://www.ivan-herman.net > > > > > > > > > > > > > > > > > > ---- > > Ivan Herman, W3C Semantic Web Activity Lead > > Home: http://www.w3.org/People/Ivan/ > > mobile: +31-641044153 > > PGP Key: http://www.ivan-herman.net/pgpkey.html > > FOAF: http://www.ivan-herman.net/foaf.rdf > > > > > > > > > > > > > > -- > -ericP -- -ericP
Received on Thursday, 17 February 2011 23:02:38 UTC