A Visual Modeling Approach for the Semantic Web Rule Language

Tracking #: 1877-3090

Authors: 
Benedikt Pittl
Hans-Georg Fill

Responsible editor: 
Thomas Lukasiewicz

<
Submission type: 
Full Paper
Abstract: 
The Semantic Web Rule Language (SWRL) is considered a main pillar for realizing the semantic web and for supporting innovative rule-based applications. Thereby, it is used to infer new knowledge from a given fact base. Today, SWRL rules are developed and managed by technical experts in text-based editors using software applications such as the Stanford Protégé toolkit. For easing the specification and analysis of SWRL rules by non-technical users, we introduce in this paper a visual modeling approach for SWRL. By building upon a visual modeling language, the approach includes validation mechanisms and layouting algorithms for visually representing new as well as existing rules. The approach further provides import and export interfaces to common SWRL exchange formats. In this way, its compatibility with widely-used reasoners and semantic web platforms is guaranteed. For ensuring its feasibility, the approach has been prototypically realized using the SeMFIS platform and evaluated using the sample rules as provided in the SWRL specification.
Full PDF Version: 
Tags: 
Reviewed

Decision/Status: 
Major Revision

Solicited Reviews:
Click to Expand/Collapse
Review #1
By Carsten Keßler submitted on 13/May/2018
Suggestion:
Major Revision
Review Comment:

This paper describes a visual approach to developing SWRL rules with the aim to make it easier for domain experts who are no comfortable writing SWRL rules as text to utilize the inference capabilities of SWRL. The paper is technically sound and details how the visual model works and how it is translated to actual SRWL rules. My main concern is that the whole premise that such a visual approach will make life easier for domain experts is not proven at all, as far as I can see. My impression is that such a domain expert will still need a fairly good understanding of the logics behind SWRL, what the use cases are, etc, so that the 'delta' in usability might be a lot smaller than the authors are trying to convince us. The main reason for my recommendation for a major revision is therefore that backing up this statement will require a usability test. Having said that, I do see that the paper is already quite extensive and dense, so an easier way to address this problem would be to simply remove this claim and introduce the visual language as it is, focus on the technical implementation, and leave the usability test (preferably with domain experts as well as 'SWRL experts') for future work.

Originality: The idea is somewhat original – not entirely new, since the authors have already proposed this approach in a prior publication, but I am not aware of similar work from any other group .

Significance of the results: As mentioned above, the authors show that their approach works from a technical point of view, i.e., they can translate between the visual model and actual SWRL rules. My main concern – i.e., that this does not help domain experts without the skills to write SWRL rules in text form too much – remains.

Quality of writing: The paper is easy to read, well organized, and written in excellent English.

Review #2
Anonymous submitted on 23/May/2018
Suggestion:
Major Revision
Review Comment:

This manuscript was submitted as 'full paper' and should be reviewed along the usual dimensions for research contributions which include (1) originality, (2) significance of the results, and (3) quality of writing.

This work develops:
1. A visual modeling approach for SWRL. By building upon a visual modeling language, the approach includes validation mechanisms and
2. Import and export interfaces to common SWRL exchange formats.
3. A prototype realized using the SeMFIS platform
The paper is moderately original because does not take into account previous work in the area and mostly follows one framework: ADOxx and FDMM .
The authors start with the claim that: “The SemanticWeb Rule Language (SWRL) is considered a main pillar for realizing the semantic web”. I think this is not entirely true. SWRL is very popular and has been used in many semantic web applications and SWRL rules are supported by the Protege OWL editor,( also in version 5.x), rule engines Jess and Drools, and reasoners Pellet and Hermit.
The authors do not mention the first SWRL metamodel based on MOF [1]. The SWRL is simple and did not make it to the W3C standards from 2004. It is most likely that it will never become a W3C standard. It was meant to be replaced by the standardized RIF. RIF was in turn too complex and did not get enough traction. The OMG continued the work on RIF by standardizing Requirements Interchange Format (ReqIF) based on CMOF; the latest version 1.2 published in 2016. It is to be seen whether this version reaches out to the industrial world.
Another development is development of SPARQL Inferencing Notation (SPIN) to represent SPARQL rules and constraints on Semantic Web models, building on the widespread acceptance of the SPARQL query language for querying and processing Linked Open Data. The successor of SPIN is Shapes Constraint Language (SHACL), the W3C Recommendation published on 20 July 2017 that would take SPIN to become a de-facto industry standard. For SWRL rule-based ontology applications to be prolonged Bassiliades [2] developed a prototype tool using SWI-Prolog that takes as input an OWL ontology with a SWRL rule base and transforms SWRL rules into SPIN rules in the same ontology, taking into consideration the object-oriented scent of SPIN, i.e. linking rules to the appropriate ontology classes as derived by analyzing the rule conditions.
Finally, the authors should take into account work on visualization of SWRL rules [3,4].

Intricate semantic differences between different rule languages and their metamodels may often be difficult to understand. The authors should more clearly state what is advantage of their SWRL metamodel based on ADOxx and Formalism for Describing ADOxx Meta Models and Models (FDMM) developed by the Austrian company BOC AG, compared to the metamodel based on CMOF.

As for competiveness of SWRL, comparison with visualization of SPIN/SHACL (supported by TopBraid TopQuadrant and AllegroGraph) should be made.

It seems to me that due to the lack of well-defined formalisms behind SWRL and because of the large amount of freedom to implement this language, SWRL remained only Submission and not the W3C standard. This work definitely better defines SWRL and is quite important.

Finally, I tried to install the SemFIS plugin from: http://austria.omilab.org/psm/content/semfis/downloadlist?view=downloads.
Unfortunately, activating the plugin gives a blank screen (tab) inside Protégé 5.x from the Stanford page. Even when I load the ontology with the SWRL rules, the plugin displays a blank card (as in the browser).
In page 2 the sentence “Four our purposes , we decided to use the Formalism for Describing ADOxx Meta Models and Models (FDMM)" should be "For our purposes…".

In summary, I am recommending to publish this work with major revisions.

[1] Saartje Brockmans, Peter Haase, Pascal Hitzler, Rudi Studer: A Metamodel and UML Profile for Rule-Extended OWL DL Ontologies. ESWC 2006: 303-316]
[2] Nick Bassiliades: SWRL2SPIN: A tool for transforming SWRL rule bases in OWL ontologies to object-oriented SPIN rules. CoRR abs/1801.09061 (2018)]
[3] Ismail Akbari, Bo Yan, Junyan Zhang, Harold Boley:
Visualizing SWRL Rules: From Unary/Binary Datalog and PSOA RuleML to Graphviz and Grailog.CSWS 2013: 56-57
[4] Jing Mei, Harold Boley:
Interpreting SWRL Rules in RDF Graphs. Electr. Notes Theor. Comput. Sci. 151(2): 53-69 (2006)