NSPARQL Jesse Weaver 20080911
From Semantic Portal Wiki
Presentation slides:
Presentation given at CSCI 6966 Advanced Semantic Web (Fall 2008) - Lesson 3
- Speaker: Jesse Weaver
- Title: nSPARQL: A Navigational Language for RDF
- Authors: Jorge Pérez, Marcelo Arenas, Claudio Gutierrez
- Conference: ISWC 2008
- URL: http://web.ing.puc.cl/~jperez/papers/nsparql_ext.pdf
- Date of Presentation: 2008/09/11
Questions
| ID | Question | Name | Answer |
|---|---|---|---|
| For NSPARQL Jesse Weaver 20080911 1 | How good is O(G) complexity (w.r.t. large amount of data)? What could the average case complexity be?
Casually: Can nSPARQL have a better representation? Isn't SPARQL (& SQL) very intuitive? A thought: Some large repositories prefer performing reasoning and storing inferred data to speed up query answering. If NSPARQL works well we can avoid some redundancy such as storing: Tom is Human, when data contains Tom is Boy. |
Ankesh Khandelwal | |
| NSPARQL Jesse Weaver 20080911 Jesse Weaver | The conclusion of time complexity O(G*exp) seems to be based on the assumption that each call to Label does some fraction of G*Di(exp) work. However, accounting for lines 5-6 of Label seems less intuitive as the proof for theorem one states only that a depth-first search could be used. Generally, a depth-first search runs in O(V+E) time for a given graph G0=(V,E). V seems like it could be O(G*exp) (exp local to Label), but how do we account for E? Furthermore, do we have to run this depth-first search for every (u,q0) in the product automaton? Am I missing something? Is there some aspect of the nature of the product automaton that I'm overlooking? | Jesse Weaver | |
| NSPARQL Jesse Weaver 20080911 Jesse Weaver 2 | See the question page. | Jesse Weaver | |
| NSPARQL Jesse Weaver 20080911 Joshua Shinavier 1 | One of the main points of this paper is that nested regular expressions can be evaluated efficiently, in linear time w.r.t. the size of the graph and to the length of the expression. From a practical point of view, though, is this really efficient? One of the nice things about traversal-based algorithms is that, depending on the topology of the graph, their performance may be more or less independent the actual size of the graph. The proposed evaluation strategy, in contrast, appears to require a massive global calculation for each step in a path expression, e.g. upwards of one billion operations for a simple one-step path from a node to its immediate neighbors in a large, billion-node graph. | Joshua Shinavier | |
| NSPARQL Jesse Weaver 20080911 Joshua Taylor 1 | Admittedly, I had a hard time following all of the mechanics of the implementation. At the end of Section 3, however, the EVAL algorithm finishes with if ... then return YES else return NO. If I made a SPARQL query of the form "Paris ?p Calais; ?p rdfs:subPropertyOf transport" I can get various values for ?p: TGV; train, and transport. Are all the values available with nSPARQL, or just the yes/no answer that there is some such ?p? | Joshua A. Taylor | |
| NSPARQL: Why address RDFS at the query level? | The nested regular expressions introduced in the paper may prove very useful to many types of queries, but no reasons are given for why one would want to address RDFS reasoning at the query level instead of the model level. Adding RDFS reasoning at the model level would allow the (standardized) query language to remain unchanged while providing the same benefits that are used in all the example queries. | Gregory Todd Williams |
| Author | Text | |
|---|---|---|
| Perez2008nsparql question 1 by lebo | Tim Lebo | The "path-like" nature resembles Fresnel Selector Language (FSL) http://www.w3.org/2005/04/fresnel-info/fsl/.
|
| Perez2008nsparql question 2 by lebo | Tim Lebo | (just a comment) The "axis-like" nature, along with the "forward-backward" directionality, resembles Ted Nelson's zzStructure http://www.nongnu.org/gzz/gi/gi.html. The zzStructure incorporates a layout that facilitates visual navigation. It would be interesting to compare this work from a different discipline in a different era (he was decades ago). |
| Perez2008nsparql question 3 by lebo | Tim Lebo | The authors note "the occurrence of variables in the predicate position of triple patterns is forbidden in nSPARQL." This limitation resembles another that was described in the DARQ paper: "The matching compares the predicate in a triple pattern with the predicate defined for a capability and evaluated the constraint for subject and object. Because matching is based on predicates, DARQ currently only supports queries with bound predicates."
|
| Perez2008nsparql question 4 by lebo | Tim Lebo |
|
Absentees
Facts about NSPARQL Jesse Weaver 20080911RDF feed
| A | Presentation +, and Presentation not attended by Tim Lebo + |
| Conference | ISWC 2008 + |
| Date | 11 September 2008 + |
| Given at | CSCI 6966 Advanced Semantic Web (Fall 2008) - Lesson 3 + |
| Paper has author | Jorge Pérez +, Marcelo Arenas +, and Claudio Gutierrez + |
| Speaker | Jesse Weaver + |
| Title of paper | nSPARQL: A Navigational Language for RDF + |
| Url | http://web.ing.puc.cl/%7Ejperez/papers/nsparql_ext.pdf + |
