Axel Polleres wrote: > Kingsley Idehen wrote: >> Stephane, >>> On behalf of Axel Polleres: >>> Kingsley, all, >>> >>> The XML/Xquery and RDF/SPARQL worlds don't need to be separate! >> Not the point I was trying to make :-) >> >> Juxtaposition in this scenario is not about mutual exclusivity in any >> form. I want to able to show the merits of both approaches en route >> to bringing clarity to an often misunderstood matter. >>> You might be interested in our XQuery+SPARQL combination called >>> XSPARQL. >>> An initial prototype and technical report are available at: >>> http://www.polleres.net/xsparql >> >> Here are some links associated with my views on SPARQL and XQuery, >> which do have a points of coherent intersection: >> >> 1. http://stage.vambenepe.com/archives/173 >> >> 2. http://weblog.infoworld.com/udell/2006/04/28.html#a1437 (2006 >> podcast interview with Jon Udell) >> > Interesting thoughts! So, what I wanted to say is: Thinking along the > same lines as you did, > we have defined the combination of both in ONE language proposal... > which might suit what you sketch here. > There is no need to keep XQuery and SPARQL separate or in a pipeline > suyntactically, they combine quite well. > > The combination solves the problem of missing features in SPARQL quite > elegantly, most of them can be done using XQuery > constructs. Also the combination is useful for all kinds of > transformaitons from RDF to XML and vice versa.
Axel, Yes, but in relation to Mark Mail the juxtaposition is ultimately about Linked Data and de-referencable URIs. The Mark Mail solution doesn't expose data items via URIs which makes the meshing of a specific data item with data in some other space tedious at best. As you know, we are trying to demonstrate how a "Web of Linked Data" is ultimately a Web in which each item of data has an HTTP based URI/ID, and via these URIs/IDs we are able to mesh disparate data items via remote data object referencing. One of my concerns and peeves is that the power and role of HTTP re. the Semantic Web picture is sometimes overlooked. Some History: Prior to today's ubiquitous Web object and object-relational databases offered what we called "Linked Data" today. The only headache, this capability was typically silo-ed by DBMS Engine (plethora of vendors and actual supported models), Development Environment and/or Frameworks (e.g NeXTStep's EOF), Query Languages (OODBMS engines agreed on OQL after the period of relevance expired), or in ability to Federate across networking boundaries. Enter HTTP and the Web: A URI is an Object ID, HTTP the messaging substrate, and today we can remotely reference Data Objects (of type Document / Information Resource and beyond) via their HTTP based URIs/ID. Also note, SPARQL is the Semantic Web's timely equivalent of OQL (note Microsoft have devised Entity SQL from this Object Query Language in their platform specific Distributed Data Object framework aka .Net's Entity Framework). Kingsley > > cheers, > Axel > >>> >>> > Hey Kingsley, >>> > >>> Tom, >>> > Nice. That would be great. Not sure how this proposal relates to >>> work by >>> > Michael H and co from last year's Semantic Web Challenge >>> > >>> > http://sw.joanneum.at/mle/xplore.php (demo) >>> > http://www.cs.vu.nl/~pmika/swc-2007/mle.pdf >>> <http://www.cs.vu.nl/%7Epmika/swc-2007/mle.pdf> (paper) >>> > >>> You mean, this may be related to work by Micheal H & Co. ? :-) >>> >>> Anyway, I see that their project is basically what I am seeking. >>> The >>> only variables would be the backend RDF store and some Front-end >>> tweaking, so that we end up with an aesthetic replica of Mark Mail. >>> >>> I think juxtaposing Mark Mail (XQuery and XML) aside a Linked Data >>> variant (SPARQL and RDF) ultimately provides great demonstration >>> and >>> tutorial material. >>> >>> Kingsley >>> > Cheers, >>> > >>> > Tom. >>> > >>> > >>> >> -----Original Message----- >>> >> From: [EMAIL PROTECTED] >>> <mailto:[EMAIL PROTECTED]> >>> >> [mailto:[EMAIL PROTECTED] >>> <mailto:[EMAIL PROTECTED]>] On Behalf >>> >> Of Kingsley Idehen >>> >> Sent: 18 March 2008 20:09 >>> >> To: [email protected] >>> <mailto:[email protected]> >> Linking Open Data >>> >> Subject: [Linking-open-data] A Potential Linked Data project >>> >> >>> >> All, >>> >> >>> >> Mark Mail <http://markmail.org/> is an interesting project >>> >> that showcases XQuery and XML. I think a Linked Data >>> >> equivalent based on SPARQL & RDF would be of great value. >>> >> >>> >> >>> >> -- >>> >> >>> >> >>> >> Regards, >>> >> >>> >> Kingsley Idehen Weblog: >>> >> http://www.openlinksw.com/blog/~kidehen >>> <http://www.openlinksw.com/blog/%7Ekidehen> >>> >> President & CEO >>> >> OpenLink Software Web: http://www.openlinksw.com >>> >> >>> >> >>> >> >>> >> >>> >> _______________________________________________ >>> >> Linking-open-data mailing list >>> >> [email protected] >>> <mailto:[email protected]> >>> >> http://simile.mit.edu/mailman/listinfo/linking-open-data >>> >> >>> >> >>> > >>> > >>> >>> >>> -- >>> >>> >>> Regards, >>> >>> Kingsley Idehen Weblog: >>> http://www.openlinksw.com/blog/~kidehen >>> <http://www.openlinksw.com/blog/%7Ekidehen> >>> President & CEO >>> OpenLink Software Web: http://www.openlinksw.com >>> >>> >>> >>> >>> _______________________________________________ >>> Linking-open-data mailing list >>> [email protected] >>> <mailto:[email protected]> >>> http://simile.mit.edu/mailman/listinfo/linking-open-data >>> >>> >> >> > > -- Regards, Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen President & CEO OpenLink Software Web: http://www.openlinksw.com _______________________________________________ Linking-open-data mailing list [email protected] http://simile.mit.edu/mailman/listinfo/linking-open-data
