Who needs ‘AS’ in SPARQL?

Lee and I were having a little brainstorm to solve an issue we have with the SPARQL Calendar Demo. Basically, we are using AJAX to query graphs sucked in by sparql.org, but because not all data on the web is stored in RDF (darn), we must make use of very helpful services to convert existing data into RDF. Now the problem lies in the fact that we would like to make use of the GRAPH keyword to “kind of” keep track who’s the author of the triples, as in (excerpt):

PREFIX foaf: <http://xmlns.com/foaf/0.1/>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX ical: <http://www.w3.org/2002/12/cal/icaltzd#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
SELECT DISTINCT ?title ?start ?end ?name ?location ?g 
WHERE {
  {
    {
      <http://torrez.us/who#elias> rdfs:seeAlso ?g .
       ?g rdf:type ical:Vcalendar .
       OPTIONAL {
        <http://torrez.us/who#elias> foaf:name ?name 
      } .
       OPTIONAL {
        <http://torrez.us/who#elias> rdfs:label ?name 
      } .

     ...snip...

     GRAPH ?g {
      _:ev ical:summary ?title; ical:dtstart ?start; ical:location ?location.
    }.
}

However, ?g is no longer what I named it inside my foaf file, instead it’s name is the url to the service with it as a parameter that converts it to RDF, for example:

http://torrez.us/elias/school.ics

versus

http://torrez.us/services/ics2rdf/http://torrez.us/elias/school.ics

Therefore wouldn’t it be nice if we could express the following in either the SPARQL Query Language or Protocol:

SELECT ?title
FROM <http://torrez.us/services/ics2rdf/http://torrez.us/elias/school.ics> AS <http://torrez.us/elias/school.ics>
WHERE { ..snip... }

I can imagine the discussion on the semantic-web mailing list over the meaning of AS and all the marriages it will destroy if not used properly.

About this entry

You’re currently reading “Who needs ‘AS’ in SPARQL?,” an entry on Elias Torres

Published:: 03.24.06 / 12am

Category:: SemWeb

About

I’m married and father of three.

I’m a Christian and worship at CBC.

I co-founded Performable.

I’m a java, python, javascript hacker.

Here’s my FOAF file (and URI).

I’m an amateur photographer.

I work on the Web.

I participate in Open Source software development (Roller, Abdera, RDFLib, WordPress).

You may contact me (email or jabber/gtalk) at .

Dan Brickley 03.24.06 / 2am

Could you be persuaded this is properly a matter for software APIs to SPARQL engines, and the language itself needn’t change?

Hmm would it be reasonable to have

FROM NAMED blah blah in the textual query, yet have the software context populate the named graph by running adaptors and filters and 3rd party services, rather than doing a simple naive HTTP GET? I believe Jena/ARQ has some facilities for allowing an alternate copy of the graph (eg. as a file: URI for disconnected use), though not sure if it is wired up to the FROM NAMED construct.

Michael 03.24.06 / 4am

Your service already uses http://torrez.us/elias/school.ics as xml:base and rdf:about=”#${id}”. I don’t see why this would be needed?

Michael 03.24.06 / 5am

Nevermind, I wrote that a few minutes after getting up, sorry

Sounds like application logic to me though, getting all that done by SPARQL itself would turn out rather messy I believe.

Andy Seaborne 03.24.06 / 5am

In Jena/ARQ, you can do that – it’s not part of the query – it’s part of the environment. There’s a mapping from name use din FROM/FROM NAMED to the real file/URL to use. Used a lot for ontology imports. It’s called a location-mapping.

Elias 03.24.06 / 8am

k. I’m persuaded that this shouldn’t be in the language and that it’s perfect place is in the software (as Jena ‘s Locator mappings). But how can I do so externally to the SPARQL engine, i.e. a web 2.0 application talking to http://sparql.org/ and the authors don’t know Andy Seaborne’s email address?

Lee Feigenbaum 03.24.06 / 9am

How can you do it? Dirty tricks, Elias, dirty tricks.

(But you knew that already

Dave Beckett 03.24.06 / 11am

As I recall in earlier DAWG discussions, we proposed FROM (NAMED )* but eventually ended up with the other 2 forms.

As with ARQ, my Rasqal code can still do this and it’s available if constructing queries via the API.

Richard Cyganiak 03.24.06 / 2pm

In line with Michael’s brainfart, I think it wouldn’t be *completely* crazy if the SPARQL processor used the xml:base of of graphs loaded via FROM NAMED as graph names. Could be confusing as hell in some situations, or just whay you want in others.

This wouldn’t violate the spec, and would be very much in line with the thinking that an RDF graph has a name (the original data source), but may be served from somewhere else (cache, converter).

Elias 03.24.06 / 2pm

Richard, I think it’s a good idea. I wonder what others think about it, would it get any traction in a best practices kind of setting?

Richard Cyganiak 03.24.06 / 8pm

I don’t know, and I’d also like to know what others think. Some wider discussion of the pragmatics around SPARQL graph names would certainly be good.

Dave Beckett 03.25.06 / 12am

There’s already a pseudo design for recording graph names in rdf/xml. The way to delimit rdf graphs is the rdf:RDF element itself, all it needs is a tiny syntax extension of an attribute such as rdf:graphURI to give the graph inside a URI name. *If* RDF/XML was worth extending.

An xml:base on the root rdf:RDF element seems an interesting way without an extension I guess it is OK.

TechnicaLee Speaking: SPARQL Calendar Demo: Retrieving Calendar Events 05.21.06 / 7am

[...] This is an ugly hack. Anyone reading this should be outraged. I was, but sometimes deadlines beckon. This is an ugly hack not least of all because the all of those ics2rdf URLs are invalid URL syntax (the reserved characters in the ical query parameter should be URL escaped). A possible solution would be to extend our calendar convention to require that a person’s FOAF data contain a triple along the lines of: <http://example/myCalendar.ics> ex:asRDF <http://torrez.us/services/ics2rdf/?ical=http%3A//example/myCalendar.ics>. That’s ugly, also, but perhaps a bit better. When we look at the queries that deal with people’s interests and make use of data from upcoming.org, we’ll see that this is a more general problem: when we access non-RDF data as RDF, how do we semantically associate the (different) URLs of (different) representations of the same data in a manner which is relatable within a SPARQL query? Elias wrote about this dilemma and solicited opinions ranging from handling this at an application level (but how do we do that with the current SPARQL Protocol over the web?) to extending the expressibility of SPARQL FROM NAMED clauses. [...]

Elias Torres

Who needs ‘AS’ in SPARQL?

About this entry

Comments are closed

Recently

Categories

Monthly Archives

About

Pages

Elias Torres

Who needs ‘AS’ in SPARQL?

About this entry

Comments are closed

Links

Blogroll

Recently

Categories

Monthly Archives

About

Pages