Peter Hansen

Department of Mathematics and Computer Science, University of Bremen

User Tools

Site Tools


“Gracefully Rewriting Into Non-recursive Datalog.”



Grind is a software tool to check for FO-rewritability of concept and conjunctive queries in EL ontologies, and to produce non-recursive datalog rewritings, if possible.

Grind is developed at the Theory of Artificial Intelligence group of University of Bremen.


Please contact the authors here.


Peter Hansen, and Carsten Lutz. Computing FO-Rewritings in EL in Practice: from Atomic to Conjunctive Queries. (Submitted to DL 2017).
Paper (PDF)BibTex entry

Peter Hansen, Carsten Lutz, İnanç Seylan, and Frank Wolter. Efficient Query Rewriting in the Description Logic EL and Beyond. In Proceedings of the 24th International Joint Conference on Artificial Intelligence (IJCAI 2015), 2015.
Paper (PDF)BibTex entry – The Grind query rewriting system

Peter Hansen, Carsten Lutz, İnanç Seylan, and Frank Wolter. Query Rewriting under EL TBoxes: Efficient Algorithms. In Proceedings of the 27th International Workshop on Description Logics (DL 2014), 2014.
Paper (PDF)BibTex entry – Test ontologies: dl2014-onto.tgz (0.5 MB)


The picture above is a derivative of "A skateboarder performing a nosegrind" by MattLaws, used under CC BY 2.0.


(The following is contained in the file README.txt distributed with Grind.)

         Grind -- Gracefully Rewriting Into Non-recursive Datalog      

                       Version 2.0 2017-05-08.

           Copyright 2014, 2015 Peter Hansen and İnanç Seylan,
                       2016, 2017 Peter Hansen.
                 This software is released under GPL 3.

Welcome to version 2.0 of Grind!


Download grind2.tar.bz2 and extract it. Then

 $ cd grind2
 $ java -jar Grind.jar
Rewrite all 50 CQs (minus aborts), log details stdout:

 $ java -jar Grind.jar


In this folder, you find a Java archive Grind.jar, and accompanying libraries
in the lib folder. Please download also the test ontologies, and expand 
them to a suitable place.

The file Grind.jar contains the source code, which can be extracted by
 $ unzip Grind.jar

This distribution of Grind 2.0 implements 50 hand-coded CQs crafted for five of
the test ontologies (ENVO, FBbi, MOHSE, not-galen, and SO). 
The queries are to be found in the file 
Run them by invoking:

 $ java -jar Grind.jar

Grind will output non-recursive datalog rewritings, if they exist.



         This software is released under GPL 3 licence.
grind.txt · Last modified: 2017/05/08 11:13 by peter

Page Tools