Ecco: A Hybrid Di Tool for OWL 2 ontologies

Abstract

The detection and presentation of changes between OWL ontologies (in the form of a di ) is an important service for ontology engineering, being an active research topic. In this paper, we present a di tool that incorporates structural and semantic techniques in order to, rstly, distinguish e ectual and ine ectual changes between ontologies and, secondly, align and categorise those changes according to their impact. Such a categorisation of changes is shown to facilitate the navigation through, and analysis of change sets. The tool is made available as a web-based application, as well as a standalone command-line tool. Both of these output an XML change set le and a transformation into HTML, which allows users to browse through and focus on those changes of utmost interest using a web-browser.

Introduction

Detecting and presenting changes between any two documents (so-called di ) is an essential service that is hardly con ned to software engineering. While regular textual di s, such as UNIX's di function, rely on the assumption that order matters, in ontologies that no longer holds; OWL [ 3 ] does not impose a systematized ordering of axioms, but instead de nes a higher level notion of syntactic equivalence (so-called structural equivalence [ 13 ], and associated notion structural di erence). In turn, this notion provides a basis for ignoring certain types of negligible changes, such as the order of axioms or concrete syntax (e.g., OWL/XML compared to RDF/XML serialisations of the same ontology).

There are a variety of di services based on structural equivalence [ 7, 12, 11 ], which distinguish additions and removals, and subsequently align axiom changes with those class names found on the left-hand side of the axiom. However, no further characterisation of changes is typically carried out, e.g., whether changes produce any logical e ect (thus e ectual ) or not (ine ectual ), as conducted in [ 11 ], or whether there exist any relations or additional properties of changes that might help users analyse and understand them. Aside from this, such di s lack a standard and essential feature of any di : an alignment between the source and target of a change. This kind of data can be collected at development time via edit-based di s, as implemented in SWOOP [ 9 ], although if there are no such change records then a post facto change analysis is impossible.

In this paper we present a di tool that incorporates structural and semantic techniques to, rstly, distinguish which additions and removals (obtained via structural di erence) are e ectual or ine ectual, and, secondly, nd the source of each change (where attainable), which in turn allows us to categorise and align (source with target of) changes between two ontologies. The categories follow from the di erent kinds of impact a change can have, e.g., by further constraining an axiom we can make it \stronger", and the relation between this stronger axiom and its preceding version is made explicit by our categorisation, and suitably presented by our tool. The latter is available either as a web-based application, or a command-line tool with more advanced features. 2

Preliminaries

We assume the reader to be reasonably familiar with ontologies and OWL, as well as the underlying description logics (DLs) [ 1 ], though detailed knowledge is not required. The di categories discussed in the paper are de ned in [ 4 ], and will be brie y discussed and exempli ed. When comparing two ontologies we refer to them as O1 and O2, and their signatures, i.e., the set of entity (class, property and individual) names occurring in them, as Oe1 and Oe2, respectively. The signature of an axiom is denoted e. A structural equivalence relation between two axioms 1 and 2 is denoted 1 s 2. Throughout this paper we use the standard description and rst order logic notion of entailment; an axiom entailed by an ontology O is denoted O j= . A justi cation J for an entailment is a -minimal subset of an ontology O that is su cient for to hold [ 8 ]. We refer to an e ectual addition (removal) from O1 to O2 as an axiom such that 2 O2 and O1 6j= ( 2 O1 and O2 6j= ) [ 4 ]. 3

Ecco: A Hybrid Di for OWL 2

We present the di tool Ecco, available on the web as a Java servlet, and accessible at http://owl.cs.manchester.ac.uk/diff. Alternatively there is also a command-line tool with more advanced features, which can be downloaded from http://owl.cs.manchester.ac.uk/research/topics/diff/. In order to demonstrate the functionality of the tool, as well as how its output can be interpreted, we start o with a comprehensive di walkthrough on toy ontologies, and further on we show the output of Ecco on those same ontologies. 3.1

Di

Walkthrough Consider ontologies O1 and O2, de ned in Table 1. From O1 and O2 we have the following structural di erences:

Additions(O1; O2) = f 1; 2; 4; 5; 6; 7; 9; 10; 11; 12; 13g Removals(O1; O2) = f 1; 3; 4; 5; 7; 8g

Note that 6 is not syntactically equal to 8 ( 6 6= 8), however they are structurally equivalent ( 6 s 8). Therefore these axioms are not reported as changes. Given the sets of structural additions and removals from O1 to O2, we check which axioms in Removals(O1; O2) are entailed by O2 (ine ectual removals), and vice-versa for Additions(O1; O2). Thus we obtain a distinction between e ectual and ine ectual changes, as follows:

E ectualAdditions(O1; O2) = f 2; 6; 10; 11; 12; 13g E ectualRemovals(O1; O2) = f 8g Ine ectualAdditions(O1; O2) = f 1; 4; 5; 7; 9g Ine ectualRemovals(O1; O2) = f 1; 3; 4; 5; 7g

There are several ine ectual changes in the change set, while e ectual changes are mostly additions (and a single removal). The changes are categorised as shown in Table 2. Consider these ine ectual additions; 9 is a rewrite of f 7; 5g, as well as an avoided redundancy (i.e., had it been added to O1 it would be redundant). The axiom is also weakened, due to 8. This may seem like an unintentional change, since now we face a loss of information regarding J , which is no longer mentioned in O2. Such a change may be worth revising. The axiom 1 is redundant, since we have from O1 that A v C, which is also entailed from O2, and O2 j= A v B. Therefore the user can dispose of this axiom.

Bear in mind that the existence of a rewritten axiom from O1 to O2 does not imply that the same holds in the opposite direction. This is applicable to all categories. Also we can have that an axiom is in more than one category. Consider axiom 1; a justi cation J1 for 1 is J1 = f 2; 3g, which indicates a strengthening (since we have that 2 2 E ectualAdditions(O1; O2)), as well as a redundancy ( 3 2 O1 \ O2). Another justi cation J2 = f 1; 3g indicates a redundancy; 1 2 Ine ectualAdditions(O1; O2), and 3 2 (O1 \ O2).

In terms of e ectual changes there is only one removal, and six additions. The e ectual removal ( 8) represents a weakening of 9 with retired terms (J is not mentioned in O2). In the analysis of the ine ectual changes it was already noted that axiom 8 should be revised. The pure additions represent adjustments to the class hierarchy, some associated with new terms in O2. Both axioms 11 and 12 are strengthenings of 4, which suggests that they could be merged, especially since there is intra-axiom redundancy. Finally there is a new term K in O2 being described via axiom 10.

Rewritten la Strengthened u t c e e n I Redundant .EWeakeningRT

Rewritten l Weakened a u t c e e n I Redundant

Strengthening

StrengtheningNT l a tu NewDescription c e PureAddition E

PureAdditionNT 3 1 4 1 3 8 9 9 7 1 7 9 11 12 10 2 6 13

Generally speaking, with such a categorisation it becomes conceivably more intuitive to navigate and understand change sets. Even though ine ectual changes are often ignored by semantic di s (ContentCVS [ 7 ], or CEX [ 10 ]), we gathered from their analysis useful information such as, e.g., that axiom 4 is strengthened in two distinct, yet partially super uous axioms ( 11 and 12). Similarly we discovered that axiom 9 is weakened, from 8, which should be reconsidered as we now have that O2 6j= F v J (J becoming a retired term). The algorithm to compute the di and its categories is straightforwardly derivable from the de nitions in [ 4 ], and heavily relies on decision procedures for entailments [ 6 ], justi cation nding [ 8 ], and module extraction algorithms [ 2 ]. Ecco is implemented in Java, based on the OWL API [ 5 ] (v3.2.4).1 The output of a di is an XML le,2 containing the axioms in the di (in Manchester syntax)3 and their respective categories. In order to present this output in a more sensible way, and allow user interaction with the output, we use XSLT4 to transform the XML le into HTML, which, together with the supplied JavaScript5 le, produces a hands-on front-end to the categorised change set.

The entry point to the di on the web is shown in Figure 1, wherein users can supply URL's for the ontologies, or browse for ontology les in the local system. Using the command-line version of Ecco, users have further optional arguments: -r Analyze the root ontologies only, not any of their imports -s Save the categorical sets as OWL ontologies -o Output the change sets to the speci ed directory -n Normalize entity URIs. I.e. if two ontologies have the same entity names in a di erent namespace, this ag establishes a common namespace -i Ignore Abox axioms -v -version Print version information and exit -h -help Print help message

The output of both implementations is the same; the XML change set with its transformation into HTML. The resulting webpage allows users to browse through the change set by focusing on general categories (e.g., additions or effectual changes only) or more speci c ones (e.g., weakenings). The webpage derived from the di example in Section 3.1 is shown in Figure 2. The basic layout displays removals on the left-hand side, in red, and additions on the right-hand side, in green. The top level links and buttons allow users to, accordingly, get the source XML le, show or hide all changes, and adjust entity rendering according to entity names, entity labels (rdfs:label), or gensyms. The latter can be used to mask entity names by replacing them with shorter symbols, which, in cases where entity names or labels are too big, reduces the amount of on-screen information, making pattern analysis easier. In the change summary we present a hierarchical structure of the categories, and the number of changes in each of them. Additionally there are help buttons to help the user understand what each category represents.

From this point onwards, users can select what kind of changes to focus on, having triggers available to inspect, and navigate to speci c categories; e.g., in Figure 3 we have focused on strengthened axioms. In this gure we see on the left-hand side the removed axioms, while on the right-hand side, this being an ine ectual change, we present the justi cations for each change (note that there 1 http://owlapi.sourceforge.net/ 2 http://www.w3.org/XML/ 3 http://www.w3.org/TR/owl2-manchester-syntax/ 4 http://www.w3.org/TR/xslt 5 https://developer.mozilla.org/en/JavaScript can be more than one justi cation, such as change with ID 6). Furthermore the tool ags those axioms that are shared between both ontologies.

Consider Figure 4, where we have focused on e ectual additions, speci cally strengthenings and new descriptions. In categories involving new or retired terms, such as new or retired descriptions, these are appropriately revealed below the axiom, as shown in Figure 4. In the case of strengthenings there is a one-to-one alignment between target and source axioms, i.e., the change and what it is a change of.

Fig. 4: Select categories of e ectual additions between O1 and O2. We have presented the di tool Ecco, and shown how its categorisation and presentation of change sets facilitates change analysis. By means of this categorisation we can group changes according to their impact, allowing users to shift their attention to speci c types of changes, rather than going through an unstructured change set while inspecting both ontologies. Moreover, with the alignment of changes between ontologies we can show the changed axioms and what they are a change of. Consequently, by analysing changes in this way, one can conceivably avoid examining the actual ontologies. Ecco is available as a web-based application, allowing users to compare ontologies with no installation necessary. When speci c requirements come into play (e.g., increased Java heap space, or tool-speci c options), the standalone version of Ecco would be most appropriate, enabling more advanced features.

References

1. Baader , F. , Calvanese , D. , McGuinness , D. , Nardi , D. , Patel-Schneider , P.F . (eds.): The Description Logic Handbook: Theory, Implementation, and Applications . Cambridge University Press ( 2003 )

2. Cuenca Grau , B. , Horrocks , I. , Kazakov , Y. , Sattler , U. : Modular reuse of ontologies: Theory and practice . J. of Arti cial Intelligence Research 31 ( 2008 )

3. Cuenca Grau , B. , Horrocks , I. , Motik , B. , Parsia , B. , Patel-Schneider , P. , Sattler , U. : OWL 2: The next step for OWL . J. of Web Semantics ( 2008 )

4. Goncalves , R.S. , Parsia , B. , Sattler , U. : Categorising logical di erences between OWL ontologies . In: Proc. of CIKM-11 ( 2011 )

5. Horridge , M. , Bechhofer , S.: The OWL API: A Java API for working with OWL 2 ontologies . In: Proc. of OWLED-09 ( 2009 )

6. Horrocks , I. , Sattler , U. : A tableaux decision procedure for SHOIQ . J. of Automated Reasoning 39 , 249 { 276 ( 2007 )

7. Jimenez-Ruiz , E. , Cuenca Grau , B. , Horrocks , I. , Berlanga Llavori , R.: Supporting concurrent ontology development: Framework, algorithms and tool . Data and Knowledge Engineering 70 ( 1 ) ( 2011 )

8. Kalyanpur , A. , Parsia , B. , Horridge , M. , Sirin , E.: Finding all justi cations of OWL DL entailments . In: Proc. of ISWC/ASWC-07 ( 2007 )

9. Kalyanpur , A. , Parsia , B. , Sirin , E. , Cuenca Grau , B. , Hendler , J.: Swoop: A Web ontology editing browser . J. of Web Semantics 4 ( 2 ) ( 2006 )

10. Konev , B. , Walther , D. , Wolter , F. : The logical di erence problem for description logic terminologies . In: IJCAR-08 . vol. 5195 ( 2008 )

11. Kremen , P. , Sm d, M., Kouba , Z. : OWLDi : A practical tool for comparison and merge of OWL ontologies . In: Proc. of DEXA-12 ( 2011 )

12. Malone , J. , Holloway , E. , Adamusiak , T. , Kapushesky , M. , Zheng , J. , Kolesnikov , N. , Zhukova , A. , Brazma , A. , Parkinson , H.E. : Modeling sample variables with an experimental factor ontology . Bioinformatics 26 ( 8 ), 1112 { 1118 ( 2010 )

13. Motik , B. , Patel-Schneider , P. , Parsia , B. : OWL 2 Web Ontology Language: Structural Speci cation and Functional-Style Syntax . W3C Cand . Rec. ( 11 June 2009 ), http://www.w3.org/TR/2009/CR-owl2 - syntax-20090611/