Conference PaperPDF Available

A Method to generate a Modular ifcOWL Ontology

Abstract and Figures

Building Information Modeling (BIM) and Semantic Web technologies are becoming more and more popular in the Architecture Engineering Construction (AEC) and Facilities Management (FM) industry to support information management , information exchange and data interoperability. One of the key integration gateways between BIM and Semantic Web is represented by the ifcOWL ontology, i.e. the Web Ontology Language (OWL) version of the IFC standard, being one of reference technical standard for AEC/FM. Previous studies have shown how a recommended ifcOWL ontology can be automatically generated by converting the IFC standard from the official EXPRESS schema. However, the resulting ifcOWL is a large monolithic ontology that presents serious limitations for real industrial applications in terms of usability and performance (i.e. querying and reasoning). Possible enhancements to reduce the complexity and the data size consist in (1) modularization of ifcOWL making it easier to use subsets of the entire ontology, and (2) rethinking the contents and structure of an ontology for AEC/FM to better fit in the semantic web scope and make its usage more efficient. The second approach can be enabled by the first one, since it would make it easier to replace some of the ifcOWL modules with new optimized ontologies for the AEC-FM industry. This paper focuses on the first approach presenting a method to automatically generate a modular ifcOWL ontology. The method aims at minimizing the dependencies between modules to better exploit the modularization. The results are compared with simpler and more straightforward solutions.
Content may be subject to copyright.
A Method to generate a Modular
ifcOWL Ontology
Walter TERKAJ a,1, and Pieter PAUWELSb
aInstitute of Industrial Technologies and Automation (ITIA-CNR), Milan, Italy
bUniversity of Ghent, Ghent, Belgium
Abstract. Building Information Modeling (BIM) and Semantic Web technologies
are becoming more and more popular in the Architecture Engineering Construction
(AEC) and Facilities Management (FM) industry to support information manage-
ment, information exchange and data interoperability. One of the key integration
gateways between BIM and Semantic Web is represented by the ifcOWL ontology,
i.e. the Web Ontology Language (OWL) version of the IFC standard, being one
of reference technical standard for AEC/FM. Previous studies have shown how a
recommended ifcOWL ontology can be automatically generated by converting the
IFC standard from the official EXPRESS schema. However, the resulting ifcOWL
is a large monolithic ontology that presents serious limitations for real industrial
applications in terms of usability and performance (i.e. querying and reasoning).
Possible enhancements to reduce the complexity and the data size consist in (1)
modularization of ifcOWL making it easier to use subsets of the entire ontology,
and (2) rethinking the contents and structure of an ontology for AEC/FM to bet-
ter fit in the semantic web scope and make its usage more efficient. The second
approach can be enabled by the first one, since it would make it easier to replace
some of the ifcOWL modules with new optimized ontologies for the AEC-FM in-
dustry. This paper focuses on the first approach presenting a method to automat-
ically generate a modular ifcOWL ontology. The method aims at minimizing the
dependencies between modules to better exploit the modularization. The results are
compared with simpler and more straight-forward solutions.
Keywords. IFC, ifcOWL, Ontology, Modularization, EXPRESS
1. Introduction
BIM (Building Information Modeling) is gaining more and more relevance in the Ar-
chitecture Engineering Construction (AEC) and Facilities Management (FM) industry to
support the digitalization of the business process. Industry Foundation Classes (IFC) [16]
is one of the standards in the BIM domain and it is widely used in industrial applications.
However, there are barriers limiting its semantic interoperability and adoption on a larger
scale [23]. Indeed, the IFC standard is provided as single schema written in EXPRESS
language [14] that is extremely large and complex, being characterized by an almost
monolithic structure. For instance, the IFC4 ADD1 EXPRESS schema contains 768 En-
1Corresponding Author: Institute of Industrial Technologies and Automation (ITIA-CNR), Milan, Italy; E-
tity data types, 206 Enumeration data types, 60 Select data types, 131 defined data types,
46 FUNCTION declarations, and 2 RULE declarations. The complex structure of IFC
jeopardizes its exploitation by industrial domains outside the core AEC applications that
may need a simple model of building, spaces, elements and their relations with geometry,
topology, monitoring, automation and control, safety, etc.
Semantic Web offers opportunities to provide more effective solutions also for the
BIM domain, by exploiting its typical enablers in terms of formal modeling language,
data distribution, extensibility, and automatic reasoning. Possible BIM solutions based
on Semantic Web technologies include:
1. an OWL version of IFC, named ifcOWL. Previous works [22,21] demonstrated
how the ifcOWL can be automatically generated by converting the IFC EX-
PRESS schema to OWL. For example, this conversion leads to an ifcOWL on-
tology [21] for IFC4 ADD1 with 1313 classes, 1580 object properties, 13867
logical axioms, and 1158 individuals.
2. the development of novel ontologies for BIM that are based on the semantic web
principles and designed exploiting modularity and extendability since the begin-
ning. Such approach is currently investigated by the World Wide Web Consor-
tium (W3C) with the Linked Building Data (LBD) Community Group that is
working on a set of loosely related ontologies for Building Topology (BOT) [24],
Product, Geometry, Automation and Control [28], etc.
Given the original complexity of the IFC schema, also the resulting ifcOWL on-
tology is considerably large and complex to load and use. The ifcOWL ontology has
many interdependencies that it becomes a huge challenge to exploit data distribution both
at Tbox and Abox level. The ontology takes full advantage of OWL2 DL expressivity
(SHIQ(D)), which can lead to a high number of assertions when handed to OWL reason-
ing engines because all axioms are loaded when the ontology is referenced by an RDF
This paper will investigate how the ifcOWL ontology can be split into separate on-
tology modules, so that end users and applications only need to select the modules that
are actually going to be used. The modularization is expected to reduce the complex-
ity and provide enablers also for future extensions and integrations. Section 2 briefly
presents related works on ontology modularization, whereas Section 3 addresses the spe-
cific problem of modularizing the ifcOWL ontology. Section 4 presents the modular-
ization algorithm and Section 5 shows the results of the application of the algorithm to
generate a modular ifcOWL ontology. Finally, conclusions are drawn in Section 6.
2. Related Works
As defined by d’Aquin et al. [10], the task of partitioning an ontology is “the process
of splitting up the set of axioms into a set of modules {M1, ..., Mk}such that each Mi
is an ontology and the union of all modules is semantically equivalent to the original
ontology O”. The topic of ontology modularization has been largely addressed in the
literature [27]. Indeed, modularity can be beneficial both during the design phase and
during the deployment and usage. Some of the benefits of modularity can be mentioned
as follows [19]:
scalability for querying data and reasoning on ontologies
scalability for evolution and maintenance
complexity management
context-awareness and personalization
Modularity can be applied to pursue different goals [7] while using different strate-
gies for modularity [19], some times also in a concurrent way:
disjoint or overlapping modules
semantics-driven strategies
structure-driven strategies (e.g. using graph decomposition algorithms)
machine learning strategies
monitoring modularization and making it evolve
Various techniques have been proposed mainly to process large ontologies and ex-
tract modules from them, e.g. [5,9,11,13,15,18]. Modularization has been applied in
various knowledge domains, for instance architectural design [6,3] and biomedical do-
main [20,26,29]. In some cases the ontologies were designed since the beginning in a
modular way, for example the TOVE ontologies [12] supporting the enterprise integra-
tion. Furthermore, when addressing a modularization problem, the definition of the eval-
uation criteria [10] plays a key role and it must be consistent with the overall goals.
3. ifcOWL Ontology and Modularity
The modularization of ifcOWL is an important step in updating the ontology so that it
can be more efficiently used in a web context. Indeed, a modular ifcOWL is expected to
performance (e.g. query and reasoning)
ease of alignment with other ontologies, also reducing overlapping
At least two strategies can be envisioned to generate a modular ifcOWL:
1. modularization by content, i.e. the definition of classes and properties are sepa-
rated based on the knowledge domain they are related to, e.g. geometry, units of
measurement, building components, HVAC, etc.
2. modularization by axiom type, i.e. separating the different axioms that are in-
cluded in ifcOWL, such as definition of classes, subsumption, data/object proper-
ties, domain/range of properties, equivalent classes, cardinality restrictions apart,
etc. This option might even align with the idea of being able to load an ifcOWL
in specific OWL profiles (OWL2 EL, OWL2 QL, OWL2 RL - see [17]). For ex-
ample, an ifcOWL version not containing cardinality restrictions could be used
to conform with the OWL2 EL profile. On the other hand, most OWL reasoners
allow to specify to which level of expressiveness (RDFS, OWL2 EL, OWL2 QL,
OWL2RL) an ontology should be loaded. Thus, when reasoning is concerned,
this first option is already supported by using an OWL reasoner with appropriate
Figure 1. The IFC data schema architecture with conceptual layers, as displayed in the introduction of the IFC
specification (IFC4 ADD1) [16]
Herein the attention is focused on the first strategy that is supported by the fact that
the IFC standard was developed in a modular way and each data type (i.e. entity, enu-
meration, select, defined) in the EXPRESS schema belongs to a specific sub-schema, as
reported in its documentation [16]. Indeed, the IFC schema consists of four layers, each
containing sub-schemas (see Figure 1) that define a part of all EXPRESS data types. For
example, the IfcActorResource schema (bottom left in Figure 1) contains 3 enumera-
tions and 8 entities. The corresponding OWL definitions could in theory be kept in a sep-
arate IfcActorResource ontology module, which would be significantly smaller than
the complete ifcOWL ontology, thus resulting in better usability. However, many of the
data types in the IFC schema are tightly interconnected with each other, not only within
a sub-schema but also between different sub-schemas. In addition, in several cases there
are reciprocal dependencies between sub-schemas, even belonging to separate layers.
For example, IfcApprovalResource imports IfcControlExtension and vice versa.
Hence, in order to make a useful modularization, a full investigation of the schema needs
to be made, and the relation between the different sub-schemas (and therefore modules)
would need to be reconsidered to a significant level and detail.
Beetz et al. 2009 [4] already proposed a modular ontology for an earlier version of
ifcOWL. The authors addressed the problem of interwoven interdependencies by mov-
ing some axioms to additional modules, named pivot ontologies, that include a set of
independent semantic clusters. However, the problem of cyclic references was not com-
pletely solved. Furthermore, the author addressed the problem of modularization of the
Abox ontologies.
4. Modularization Algorithm
The proposed modularization algorithm can be applied to convert any EXPRESS schema
(e.g. IFC [16], but also ISO 15531, ISO 14649, etc.) to a modular OWL ontology. A novel
algorithm was developed to exploit the peculiar problem settings, since the modulariza-
tion takes place while converting an EXPRESS schema (e.g. IFC schema) to an OWL
ontology (e.g. ifcOWL), instead of being executed on an already existing large mono-
lithic ontology (e.g. the already generated full ifcOWL). Once the algorithm assigns an
EXPRESS definition to a module, then the conversion to the corresponding OWL axiom
is executed as stated in [21] and described with more details in [22]. It must be noted that
an automatic conversion of a technical standard from an EXPRESS schema to an OWL
ontology may lead to problems related to the lack of a precise definition and meaning of
some concepts, thus hindering semantic interoperability. Therefore, a proper ontological
analysis of the original standard should be carried out, as addressed in the works [2,25,8].
However, such analysis goes beyond the scope of this work.
The goal of the algorithm consists in finding the best way of implementing a given
input modularization by minimizing the number of direct import relations between mod-
ules. Moreover, the algorithm must avoid to create reciprocal dependencies between
modules because it would lead to circular import paths. Even though circular import
is not forbidden according to OWL2, still it is not desirable because it would actually
weaken the modularization. Indeed, a direct import of any node in a circular path will
lead to indirectly importing all the nodes in the circle; thus the final effect is that all the
modules in a circular path are merged.
In summary, the algorithm receives as input the following pieces of information:
content of a parsed EXPRESS schema in terms of data types (i.e. defined data, en-
tity, select, enumeration), subsumption relationships and attributes of each entity
data type.
input modularization in terms of mapping between EXPRESS data types and
modules. This mapping can be the results of more or less sophisticated method-
ologies, or it can be provided in a technical documentation (as in the case of
IFC [16]), or it can be simply set by the user based on his/her needs.
priority level associated with each module. This priority is used to set import rela-
tions between modules. Ceteris paribus, the module with lower priority will im-
port the module with higher priority. For instance, the priority may be associated
with the layer in the whole IFC schema, giving highest priority to the modules in
the Resource Layer and the lowest to the modules in Domain Layer.
The modularization algorithm is decomposed into two routines Algorithm 1 and Al-
gorithm 2. Algorithm 1, via the function GenModularETO, elaborates the various EX-
PRESS definitions that must be converted to a corresponding OWL axiom. The OWL
axiom is serialized as a set of triples that are added to a specific module based on the
result of the function SetModule in Algorithm 2. Thus, the function SetModule incre-
mentally adds import relationships between modules based on the actual needs derived
from the inter-module dependencies between EXPRESS data types. After STEP 4 of Al-
gorithm 1 all the OWL axioms required to convert the EXPRESS schema are assigned to
a specific module. Moreover, the full set of dependencies (i.e. import relations involving
the term owl:import) between modules is available and can be represented as a directed
graph, where the modules are nodes and the import relations are arcs. With reference to
the notation adopted in the algorithm, the graph can be defined as G= (M,I), where M
is the set of modules (i.e. nodes) and Iis the set of direct import relations (i.e. arcs). If
(w,z)I, then it means that module wMdirectly imports module zM.
Algorithm 1 Modularization Algorithm
Input: set Ent of EXPRESS entities
set Enu of EXPRESS enumerations
set Sof EXPRESS selects
set Dof EXPRESS defined data types
set of supertypes sup(t)of EXPRESS data type t(Ent Enu SD)
set of items it(s)belonging to the EXPRESS select sS
set attr(e)of attributes of entity eEnt
data type ran(e,a)(Ent E nu SD)being the range of attribute aattr(e)
set Mof modules
module mod(t)Mto which the data type t(Ent E nu SD)is assigned
set Iof ordered pairs of modules defining direct import relations
function GEN MODULA RE TO(E nt,E nu,S,D,sup,it,att r,ran,M,mod)
for all t(Ent E nu SD)do .STEP 1
add the OWL axiom defining cto module mod(t)
for all t(Ent E nu SD)do .STEP 2
for all asup(t)do
add the OWL axiom defining the subsumption realtionship to the module returned
by SE TMODULE(mod(t),mod(a),I)
for all sSdo .STEP 3
for all ait(s)do
add the OWL axiom defining the subsumption relation between sand ato the
module returned by SE TMODULE(mod(s),mod(a),I)
for all eEnt do .STEP 4
for all aattr(e)do
add the OWL axiom defining the attribute relation (i.e. property defini-
tion and restrictions) between eand ran(e,a)to the module returned by
SET MODULE(mod(e),mod(ran(e,a)),I)
Apply the transitive reduction to the graph G= (M,I).STEP 5
The result of Algorithm 1) and 2 is a directed acyclic graph (DAG), i.e. cycles in the
graph are avoided. It can be demonstrated that the resulting graph is a DAG by consider-
ing that a topological ordering is possible if and only if the graph has no directed cycles.
A topological ordering can be generated from the resulting graph because each pair of
nodes (i.e. modules) can be ordered, since Algorithm 2 guarantees that there is only one
import direction (direct or indirect) between them. Axioms involving atoms belonging to
two different modules are added always to the same module, thus solving the problem of
circular imports without needing to merge modules.
The resulting graph can be further optimized by applying a transitive reduction [1]
that allows to obtain a graph with fewer arcs but the same reachability (cf. STEP 5 of Al-
Algorithm 2 Set Module Algorithm
Input: set Mof modules
priority p(m)of module mM
set Iof ordered pairs of modules defining direct import relations
Output: selected module
updated set I
function SET MODULE(x,y,I)
if x=ythen
return x
Calculate the transitive closure of graph G= (M,I)to obtain the set of reachability
relations R
if (x,y)Rthen
return x
else if (y,x)Rthen
return y
else if p(x)>p(y)then
add (y,x)to the set I,return y
add (x,y)to the set I,return x
gorithm 1). In case of a DAG the transitive reduction is unique and consists in a subgraph
of the original graph that minimizes the number of arcs, i.e. the number of the imports.
5. Experiments
This section presents the experiments related to the generation of a modular ifcOWL on-
tology from the IFC4 EXPRESS schema2. As reported in Table 1, the input modulariza-
tion is based on the 38 IFC sub-schemas (Figure 1), plus the ontology modules express3
and list4that are automatically included during the EXPRESS to OWL conversion.
Table 1 reports also the priority level associated with each module, as required to execute
the algorithm. Three different versions of modularization algorithm have been tested to
demonstrate the benefits of the full version presented in Section 4:
1. Simple version, i.e. the modularization algorithm consisting of Algorithm 1 and
Algorithm 3 that represents a simplification of Algorithm 2.
2. Basic version, the modularization algorithm consisting of Algorithm 1 and Algo-
rithm 2, but without STEP 5 in Algorithm 1.
3. Full version, i.e. the modularization algorithm consisting of Algorithm 1 and
Algorithm 2.
Algorithm 3, used in the Simple version, implements the selection of the module
where the OWL axioms are added by looking at the incumbent need, without considering
the already set module dependencies. This simplification leads to a higher number of
direct import relations (189) compared to the Basic version (95). Moreover, the Simple
version causes the realization of circular import patterns (e.g. modules 10 and 11 import
each other), thus disabling the chance to execute a straightforward and deterministic
transitive reduction.
Table 1. Modules of the ifcOWL ontology with definition of id and priority level.
Module IFC Layer Label Priority
list N/A 1 5
express N/A 2 5
Algorithm 3 Simple version of Set Module Algorithm
Input: set Mof modules
priority p(m)of module mM
set Iof ordered pairs of modules defining direct import relations
Output: selected module
updated set I
1: function SET MODULE(x,y,I)
2: if (x,y)/Ithen
3: add (x,y)to the set I
4: return x
The comparison between the Basic and Full versions show the impact of the transi-
tive reduction, since the total number of import relations becomes 46. A synthetic com-
parison of the three algorithm versions is reported in Table 2, showing that a great deal
of unnecessary imports can be eliminated. The graph-based representation of the three
modular ifcOWL solutions are shown in Figures 2 and 3 highlighting how strongly in-
terconnected are the IFC sub-schemas. Analyzing Figure 3b, it can be noted that:
there are modules in the Core layer (i.e. IfcControlExtension and IfcProcessEx-
tension) that are not actually used in any of the modules in the upper levels;
just two modules in the Resource layer are not imported (directly or indirectly)
by IfcKernel, i.e. IfcStructuralLoadResource and IfcMaterialResource;
just two modules in the Interoperability layer are imported by modules in the Do-
main layer, i.e.
Table 2. Synthetic results of the three versions of modularization algorithm for the ifcOWL ontology, consid-
ering modules 3-40 defined in Table 1
Simple Basic Full
n. modules 38 38 38
total n. imports 189 96 47
max n. imports per module 12 6 2
avg n. imports per module 4.97 2.5 1.21
circular imports yes no no
The experiments demonstrate how the proposed algorithm enables to optimize the
number of import relations. This result is important because it leads to a decomposed
ifcOWL ontology with a minimal number of inter-dependencies, thus easing the selection
and extraction of a subset of modules that may better fit the requirements of a user.
Finally, even if the same Full version of the algorithm is adopted, different solutions
can be obtained based on the priority assigned to the modules. Indeed, the final result of
the algorithm is deterministic only if all modules have a different priority. On the other
hand, if an import relation is required between two modules having the same priority,
then the direction of the import depends on the order of the definitions that are elaborated
by Algorithm 1. For example, since modules 14 and 15 need each other (see Figure 2) and
have the same priority, the final solution could include that module 15 imports module 14,
instead of vice versa as in the experiment (see Figures 3a and 3b).
6. Conclusions
This paper presented an approach to generate a modular version of an OWL ontology
that is automatically converted from an EXPRESS schema. The attention was focused
on the case of the IFC schema given the large size of the resulting ifcOWL ontology. A
modular version of ifcOWL can help to solve practical problems related to its usability
and the scalability of software applications based on it. Moreover, the modularization
algorithm can be used also to extract fragments of the ifcOWL that are relevant for the
specific applications. This can be achieved by missing to assign some EXPRESS data
types to any module. Further developments will address:
the generation of additional OWL modules to convert the IFC Property Sets that
are currently not included in the IFC EXPRESS schema
the investigation of other modularization strategies, e.g. the second one presented
in Section 3, and the introduction of criteria to at least partially control the defi-
nition of dependencies between modules, e.g. by optimizing their priorities
testing the benefits of working with a subset of ifcOWL modules from a compu-
tational perspectives
Figure 2. Modular ifcOWL ontology resulting from the Simple version of the algorithm. The labels of the
nodes are defined in Table 1
the integration of a fragment of the ifcOWL ontology with other ontologies
the comparison of the modular ifcOWL with other ontologies for BIM that are
designed to be modular since the beginning [28]
modularization strategies for Abox ontologies [4].
This work has been partially funded by the Italian research project Smart Manufactur-
ing 2020 within the Cluster Tecnologico Nazionale Fabbrica Intelligente.
[1] A. V. Aho, M. R. Garey, and J. D. Ullman. The transitive reduction of a directed graph. SIAM Journal
on Computing, 1(2):131–137, 1972.
(a) Basic version (b) Full version
Figure 3. Modular ifcOWL ontology resulting from the Basic (a) and Full (b) versions of the algorithm. The
labels of the nodes are defined in Table 1
[2] P. P. F. Barcelos, G. Guizzardi, A. S. Garcia, and M. E. Monteiro. Ontological evaluation of the itu-t
recommendation g.805. In 2011 18th International Conference on Telecommunications, pages 232–237,
May 2011.
[3] J. Bateman, S. Borgo, K. L¨
uttich, C. Masolo, and T. Mossakowski. Ontological modularity and spatial
diversity. Spatial Cognition and Computation, 7(1):97–128, 2007.
[4] J. Beetz, J. van Leeuwen, and B. de Vries. Ifcowl: A case of transforming express schemas into ontolo-
gies. Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 23(1):89101, 2009.
[5] C. Bezerra, F. Freitas, J. Euzenat, and A. Zimmermann. ModOnto: A tool for modularizing ontologies.
In Proc. 3rd workshop on ontologies and their applications (Wonto), page No pagination., Salvador de
Bahia, Brazil, Oct. 2008. bezerra2008a.
[6] M. Bhatt, J. Hois, and O. Kutz. Ontological modelling of form and function for architectural design.
Applied Ontology, 7(3):233–267, 2012.
[7] S. Borgo. Goals of modularity: A voice from the foundational viewpoint. In WoMO, pages 1–6, 2011.
[8] S. Borgo, E. M. Sanfilippo, A. ˇ
c, and W. Terkaj. Ontological Analysis and Engineering Standards:
An Initial Study of IFC, pages 17–43. Springer International Publishing, Cham, 2015.
[9] B. Cuenca Grau, I. Horrocks, Y. Kazakov, and U. Sattler. Extracting modules from ontologies: A logic-
based approach. In H. Stuckenschmidt, C. Parent, and S. Spaccapietra, editors, Modular Ontologies:
Concepts, Theories and Techniques for Knowledge Modularization, pages 159–186, Berlin, Heidelberg,
2009. Springer Berlin Heidelberg.
[10] M. d’Aquin, A. Schlicht, H. Stuckenschmidt, and M. Sabou. Criteria and evaluation for ontology mod-
ularization techniques. In H. Stuckenschmidt, C. Parent, and S. Spaccapietra, editors, Modular Ontolo-
gies: Concepts, Theories and Techniques for Knowledge Modularization, pages 67–89, Berlin, Heidel-
berg, 2009. Springer Berlin Heidelberg.
[11] P. Doran, V. Tamma, and L. Iannone. Ontology module extraction for ontology reuse: An ontology
engineering perspective. In Proceedings of the Sixteenth ACM Conference on Conference on Information
and Knowledge Management, CIKM ’07, pages 61–70, New York, NY, USA, 2007. ACM.
[12] M. S. Fox and M. Gruninger. Enterprise modeling. AI magazine, 19(3):109, 1998.
[13] B. C. Grau, I. Horrocks, Y. Kazakov, and U. Sattler. Just the right amount: Extracting modules from
ontologies. In Proceedings of the 16th International Conference on World Wide Web, WWW ’07, pages
717–726, New York, NY, USA, 2007. ACM.
[14] International Organization for Standardization. ISO 10303-11: Industrial automation systems and in-
tegration - Product data representation and exchange - Part 11: Description methods: The EXPRESS
language reference manual, 2004. Available online: catalogue/catalogue tc/-
catalogue detail.htm?csnumber=38047 (Last accessed on 14 September 2017).
[15] R. Kontchakov, L. Pulina, U. Sattler, T. Schneider, P. Selmer, F. Wolter, and M. Zakharyaschev. Minimal
module extraction from dl-lite ontologies using qbf solvers. In Proceedings of the 21st International
Jont Conference on Artifical Intelligence, IJCAI’09, pages 836–841, San Francisco, CA, USA, 2009.
Morgan Kaufmann Publishers Inc.
[16] T. Liebich, Y. Adachi, J. Forester, J. Hyvarinen, S. Richter, T. Chipman, M. Weise, and J. Wix. Industry
Foundation Classes IFC4 official release, 2013. Available online:
ifc/IFC4/final/html/index.htm. Last accessed on 14 January 2015.
[17] B. Motik, B. C. Grau, I. Horrocks, Z. Wu, A. Fokoue, and C. Lutz. Owl 2 web ontology lan-
guage profiles (second edition), W3C Recommendation 11 December 2012. Available online: (Last accessed on 12 July 2017).
[18] S. Oh and H. Y. Yeom. Evaluation criteria ontology modularization tools. In 2011 IEEE/WIC/ACM
International Conferences on Web Intelligence and Intelligent Agent Technology, volume 3, pages 365–
368, Aug 2011.
[19] C. Parent and S. Spaccapietra. An overview of modularity. In H. Stuckenschmidt, C. Parent, and S. Spac-
capietra, editors, Modular Ontologies: Concepts, Theories and Techniques for Knowledge Modulariza-
tion, pages 5–23, Berlin, Heidelberg, 2009. Springer Berlin Heidelberg.
[20] J. Pathak, T. M. Johnson, and C. G. Chute. Survey of modular ontology techniques and their applications
in the biomedical domain. Integr. Comput.-Aided Eng., 16(3):225–242, Aug. 2009.
[21] P. Pauwels, T. Krijnen, W. Terkaj, and J. Beetz. Enhancing the ifcowl ontology with an alternative
representation for geometric data. Automation in Construction, 80:77 – 94, 2017.
[22] P. Pauwels and W. Terkaj. EXPRESS to OWL for construction industry: Towards a recommendable and
usable ifcOWL ontology. Automation in Construction, 63:100–133, 2016.
[23] P. Pauwels, S. Zhang, and Y.-C. Lee. Semantic web technologies in aec industry: A literature overview.
Automation in Construction, 73(Supplement C):145 – 165, 2017.
[24] M. H. Rasmussen, P. Pauwels, C. A. Hvid, and J. Karlshøj. Proposing a Central AEC Ontology that
allows for Domain Specific Extensions. In LC3 2017: Volume I Proceedings of the Joint Conference on
Computing in Construction (JC3), pages 237–244, Heraklion, Greece, July 2017.
[25] F. Ruy, R. Falbo, M. Barcellos, and G. Guizzardi. An ontological analysis of the iso/iec 24744 meta-
model. Frontiers in Artificial Intelligence and Applications, 267:330–343, 2014. cited By 4.
[26] A. Sojic, W. Terkaj, G. Contini, and M. Sacco. Modularising ontology and designing inference pat-
terns to personalise health condition assessment: the case of obesity. Journal of Biomedical Semantics,
7(1):12, May 2016.
[27] H. Stuckenschmidt, C. Parent, and S. Spaccapietra. Modular ontologies: concepts, theories and tech-
niques for knowledge modularization, volume 5445. Springer, 2009.
[28] W. Terkaj, G. F. Schneider, and P. Pauwels. Reusing domain ontologies in linked building data: the
case of building automation and control. In Proceedings of the 8th International Workshop on Formal
Ontologies Meet Industry, 2017.
[29] P. Wennerberg, K. Schulz, and P. Buitelaar. Ontology modularization to improve semantic medical
image annotation. Journal of Biomedical Informatics, 44(1):155 – 162, 2011. Ontologies for Clinical
and Translational Research.
... The first type of ontologies includes the Basic Formal Ontology (BFO) [46], the PROV Ontology (PROV-O) [47], the Friend of a Friend (FOAF) ontology [48], the Organization Ontology (ORG) [49], ifcOWL [50,51], and the Building Topology Ontology (BOT) [52]. BFO is an upper-level ontology that defines the fundamental categories and their relations in order to support information integration and may be used to provide top-level terms to CW entities and relations. ...
... In Fig. 12, related ontologies for different information sources are mapped to corresponding classes in the DiCon with different types of relations. For example, the building element (ifc:IfcElement and bot:Element) and spatial element classes (ifc:IfcSpatialElement and bot: Zone) in both the IFC/ifcOWL [50,51] and BOT [52] ontologies are respectively defined as subclasses of the BuildingObject or Location in the DiCon. Such linkages provide a portal to link CW with data from BIM or Linked Building Data (LBD) to use their product or spatial information, or for further related information. ...
Full-text available
With ongoing advancements in information and communication technologies (ICTs) in all stages of the construction lifecycle, information from entities related to construction workflow (CW) can now be automatically collected. These implementations are point solutions, which require systematic integration to combine their information to enable a holistic picture of CW. The major barrier to such integration is information heteroge-neity, where the information is collected from different systems under multiple contexts. Scholars in the construction domain have explored the use of ontology to solve the information-integration problem, although an ontology that both adequately represents the CW and integrates the digitalized information of CW via various systems and multiple contexts is currently missing from the existing literature. This research thus presents an ontology set for formalizing and integrating CW information within the digital construction context. The proposed digital construction ontologies (DiCon) are shared representations of construction domain knowledge that specify the terms and relations of CWs and their related information. We developed the DiCon based on a hybrid ontology development approach. The DiCon includes six modules: Entities, Processes, Information, Agents, Variables, and Contexts. The developed DiCon was further evaluated by approaches including automatic consistency checking, criteria-based evaluation, expert workshops, and task-based evaluation and involved two use cases by answering relevant competency questions via SPARQL queries. The results of the evaluation demonstrate that the DiCon ontologies are sufficient to represent domain knowledge and can formalize and integrate CW information within the digital construction context.
... Some of the most popular ontologies associated with dynamic data types, such as smart building operations and smart device integration, are BRICK and SAREF (BrickSchema, 2016;ETSI, 2013). Examples, of ontologies associated with static data types are BOT (Rasmussen, 2021), which describes topological relationships between building components and is aligned with building operations ontologies, and ifcOWL (Terkaj & Pauwels, 2017), which represents the IFC schema for building construction data in the web ontology language (OWL) (OWL, 2012). A limited number of ontologies focus on performance analytics, such as SimModelRDF (Pauwels, Corry, O'Donnell, 2014), which is an OWL translation of the SimModel XML schema and Performance Framework (PF), a high-level description of the relations between performance metrics and performance objectives (Corry et al. 2015). ...
... OBO describes building objects, however, it addresses the usability and over-specification issues of ontologies, such as ifcOWL, which lacks hierarchical structure that challenges implementation. This issue has been previously identified by Tejkaj , as well as a need for a new ontology to cover formal building object information (Terkaj & Pauwels, 2017). ifcOWL is, while a robust representation, closely aligned to, and incapable of capturing concepts not present in IFC. ...
Conference Paper
Full-text available
While the Architecture Engineering Construction and Owner-operated (AECO) industry has been successful in digitizing data concerning buildings through Building Information Modeling (BIM) applications, transforming these data into usable digital services (digitalizing) has not been fully addressed. The Semantic Web allows for the creation of abstraction layers that enable building data as a service. This paper proposes Semantic Web ontologies for representing buildings, the relationships between their elements and analytical data, along with attendant annotation systems. This method enables bi-directional exchanges between heterogeneous platforms, introducing flexibility in representing, sharing and re-using data. The work demonstrates a framework for the digitalization of building data and a service-oriented model, improving stakeholder collaboration.
... Modularising the IFC could, hence, simplify several problems that currently occur, such as schema overhead when describing small projects or singular products, filtering large data amounts, or certifying software applications for the entire schema. This is not a new idea, Terkaj and Pauwels [163] already proposed a modularisation based on the RDF serialisation of IFC, the ifcOWL. Nonetheless, it should be noted that the IFC are intended as an exchange schema for digital building models and not product descriptions. ...
... While one of their topics is the Semantic Web and Linked Data, they do not restrict to those technologies but have a broader spectrum of interests, such as the Web of Things and telecommunication in general. xx,28,38,43,50,51,156,[160][161][162][163] ...
... These federated queries -queries that can be distributed amongst several linked data sources -are submitted to a single authority, which responds by querying the distributed members of the database and returning an aggregation of those results. The approach is beneficial because data can be maintained in its original format, though some compatible API must be provided to make each source queryable [43,44]. LBD also admits the concept of a common context; in this approach, relationships between represented entities are captured through relationships that are defined as part of some ontology defined using the Web Ontology Language (OWL). ...
Interactions between built infrastructure are complex and nuanced; changes to any one component can have disproportionate effects on the system as a whole. For instance, adoption of heat pumps or electric vehicles by a significant proportion of a population in an urban centre would place new demands on both electricity transmission and distribution networks. It is essential that planners – both national and local – can understand and share information about the resource demands that this type of change places on national and local infrastructure. Access to integrated sources of information – from building component to national levels – is key to supporting policy makers and decision takers. However, over time, information – and as a consequence, the software that manages it – has evolved into functional silos; this has, in turn, affected the definition of data exchange standards. This limits the ability of experts in functional areas to exchange data and implement broader decision support systems. This paper describes the use of linked data approaches to permit queries across large, diverse information sources to provide reasoning about complex questions at multiple scales. The methodology defines a central context to which various external sources can be attached. These distributed sources are, in themselves, registered in a central catalogue; they remain, however, under the control of their source organisations. In this way a large, extensible, interconnected network of distributed data describing, for example, a built environment or electricity transmission network; this network of data resources can be queried centrally to provide customised views of subsets of the data, and so provide a richer view than one source in isolation. The approach was applied to prepare and integrate information about Ireland’s transmission grid and administrative boundaries, along with domestic housing stock into a single data source. The resulting data network is queried by a scenario exploration tool. This tool successfully allows analysis, at a national level by economists, of the effects of the adoption of new technologies on the national grid of Ireland.
... The process results in a very rich ontology (1326 classes, 1596 properties, 1162 named individuals), which can pose issues related to its management, reuse, and understandability. Aware of these and other "disadvantages", Terkaj and Pauwels [39] proposed a modularization approach to the ifcOWL ontology. The conversion of the H-BIM IFC model to IfcOWL will be explored in the present paper. ...
Full-text available
Intervention projects for historical buildings depend on the quality of multidisciplinary data sets; their collection, structure, and semantics. Building information model (BIM) based workflows for historical buildings accumulate some of the data sets in a shared information model that contains the building’s geometry assemblies with associated attributes (such as material). A BIM model of any building can be a source of data for different engineering assessments, for example, solar and wind exposure and seismic vulnerability, but for historic buildings it is particularly important for interventions like conservation, rehabilitation, and improvements such as refurbishment and retrofitting. When the BIM model is abstracted to a semantic model, enabling the use of semantic technologies such as reasoning and querying, semantic links can be established to other historical contexts. The semantic technologies help historic building experts to aggregate data into meaningful form. Ontologies provide them with an accurate knowledge representation of the concepts, relationships, and rules related to the historic building. In the paper, we are proposing an improved workflow for the transformation of a heritage BIM model to a semantic model. In the BIM part the workflow demonstrates how the fully parametric modelling of historical building components is relevant, for example, in terms of reusability and adaptation to a different context. In the semantic model part, ontology reuse, reasoning, and querying mechanisms are applied to validate the usability of the proposed workflow. The presented work will improve knowledge-sharing and reuse among stakeholders involved in historic building projects.
... The schema level mediation aims to reconstruct the model structure of ifcOWL and then apply the modular ifcOWL on a specific purpose. The concept of modular ifcOWL was firstly proposed by Terkaj & Pauwels [32] and has been adopted in several types of research, such as the Building Topology Ontology [31] and ifcOWL-DfMA Ontology [33]. The instances level mediation is intended to map the concepts with similar semantics between ifcOWL and regulatory ontology by a set of mapping rules. ...
Code compliance checking plays a critical role that identifies substandard designs according to regulatory documents and promises the accuracy of the designs before construction. However, the traditional code compliance checking process relies heavily on human work. To help the users better understand the checking process, this study proposes a gray-box checking technique and a BIM-based (Building Information Modeling) automated code compliance checking methodology that leverages ontology. The proposed approach contains a code ontology, a designed model ontology, a merged ontology, a code compliance checking ontology, a set of mapping rules, and a set of checking rules. During the checking process, the ontologies provide knowledge bases, and the rules provide necessary logic. A five-step roadmap is proposed for code ontology development for domain experts. For the time being, pre-processing is applied to create the designed model ontology to achieve time saving. Next, an ontology mapping procedure between the code and the designed model ontology is executed to obtain the merged ontology. In the ontology mapping procedure, the mapping rules aim to mitigate the semantic ambiguity between design information and regulatory information and enrich building information's semantics. Subsequently, rule-based reasoning is applied based on the checking rules and the merged ontology for checking reports generation. Finally, according to Chinese building codes, an automated code compliance checking platform is implemented for real construction projects to validate the proposed methodology.
... However, within the RDF serialisation of IFC, ifcOWL, some of the IFC's issues can be handled. Yet the criticism of the ifcOWL regarding its lack of modularity and verbose modelling of properties due to the literal translation of the IFC schema remainFarias et al. (2015), Terkaj and Pauwels (2017), Pauwels and Roxin (2016). Hence, the ifcOWL does not comply with the presented considerations. ...
The digitalisation of the Architecture, Engineering and Construction domain introduced new methods for digital collaboration, i.e. Building Information Modelling (BIM). While this method focuses on building data, the distribution of digital product models is still problematic, complicating uniform product searches and automated product data processing. Existing schemas, such as a subpart of the Industry Foundation Classes or the German VDI 3805, rely on rigid or template-driven schemas, that do not support the description of innovative or multi-functional products or impose a large schema overhead and complexity on manufacturers. Therefore, this article combines flexible and modular product descriptions with Semantic Web technologies and Linked Data. By applying Web-based technologies, the searchability of product data and the applicability of distributed data are expected to be enhanced. More precisely, this article proposes a concept for Linked Building Product Data and introduces the generic Building Product Ontology as a potential core schema of the concept. To demonstrate the feasibility of Linked Building Product Data and the Building Product Ontology, the authors apply both the concept and the data schema to innovative and multi-functional example products that cannot be described with the existing approaches for product descriptions. The evaluation demonstrates the flexibility, modularity and overall suitability of the presented concepts, meeting all collected requirements for digital product descriptions. Hence, Linked Building Product Data may solve existing issues with rigid product description schemas. At the same time, this approach complements the current research trend of Linked Building Data.
Building Information Modelling (BIM) is a process for managing construction project information in such a way as to provide a basis for enhanced decision-making and for collaboration in a construction supply chain. One impediment to the uptake of BIM is the limited interoperability of different BIM systems. To overcome this problem, a set of Industry Foundation Classes (IFC) has been proposed as a standard for the construction industry. Building on IFC, the ifcOWL ontology was developed in order to facilitate representation of building data in a consistent fashion across the Web by using the Web Ontology Language (OWL). This study presents a critical analysis of the ifcOWL ontology and of the associated interoperability issues. It shows how these issues can be resolved by using Basic Formal Ontology (ISO/IEC 21838-2) as top-level architecture. A set of competency questions is used as the basis for comparison of the original ifcOWL with the enhanced ontology, and the latter is used to align with a second ontology – the ontology for building intelligent environments (DOGONT) – in order to demonstrate the added value derived from BFO by showing how querying the enhanced ifcOWL yields useful additional information.
Full-text available
Purpose Facilitating the information exchange and interoperability between stakeholders during the life-cycle of an asset can be one of the fundamental necessities for developing an enhanced information exchange framework. Such a framework can also improve the successful accomplishment of building projects. This paper aims to use Semantic Web technologies for facilitating information exchange within existing building projects. Design/methodology/approach In real-world building projects, the construction industry’s information supply chain may initiate from near scratch when new building projects are started resulting in diverse data structures represented in unstructured data sources, like Excel spreadsheets and documents. Large-scale data generated throughout a building's life-cycle requires exchanging and processing during an asset's Operation and Maintenance (O&M) phase. Building information modelling (BIM) processes and related technologies can address some of the challenges and limitations of information exchange and interoperability within new building projects. However, the use of BIM in existing and retrofit assets has been hampered by the challenges surrounding the limitations of existing technologies. Findings The aim of this paper is twofold. Firstly, it briefly outlines the framework previously developed for generating semantically enriched 3D retrofit models. Secondly, a framework is proposed focussing on facilitating the information exchange and interoperability for existing buildings. Semantic Web technologies and standards, such as Web Ontology Language and existing AEC domain ontologies are used to enhance and improve the proposed framework. Originality/value The proposed framework is evaluated by implementing an example application and the Resource Description Framework data produced by the previously developed framework. The proposed approach makes a valuable contribution to the asset/facilities management (AM/FM) domain. It should be of interest to various FM practices for existing assets, such as the building information/knowledge management for design, construction and O&M stages of an asset’s life-cycle.
Full-text available
Structured representations of phenomena from the real world in a digital geospatial environment are essential for developing, maintaining, and using the built and natural environment. In the real world, the phenomena relate to, influence and are influenced by other phenomena through their location, shape and extent. These geospatial characteristics and relations are vital in a digital environment as well. The research presented in the thesis has studied technologies for modelling geospatial information in the three application domains of Geographic Information Systems (GIS), Intelligent Transport Systems (ITS) and Building Information Modelling (BIM). The three application domains have distinct roles in a digital geospatial environment but describe and handle many of the same real-world phenomena. Therefore, exchange and reuse of information between application domains, life cycle stages and stakeholders should be possible. The research showed that improved syntactic interoperability could be achieved by describing information models from all three application domains according to a joint approach for information modelling. Improved semantic interoperability could be achieved by using the same core concepts in distinct information models. However, a complete harmonization of information models would not be appropriate, as information models from the three application domains need to describe the real world in different contexts. Therefore, Semantic Web technologies for linking and mapping should be applied for further improvements of semantic interoperability.
Conference Paper
Full-text available
Linked data and semantic web technologies are gaining impact and importance in the Architecture, Engineering, Construction and Facility Management (AEC/FM) industry. Whereas we have seen a strong technological shift with the emergence of Building Information Modeling (BIM) tools, this second technological shift to the exchange and management of building data over the web might be even stronger than the first one. In order to make this a success, the AEC/FM industry will need strong and appropriate ontologies, as they will allow industry practitioners to structure their data in a commonly agreed format and exchange the data. Herein, we look at the ontologies that are emerging in the area of Building Automation and Control Systems (BACS). We propose a BACS ontology in strong alignment with existing ontologies and evaluate how it can be used for capturing automation and control systems of a building by modeling a use case.
Conference Paper
Full-text available
In the last years, several ontologies focused on structuring domain specific information within the scope of Architecture, Engineering and Construction (AEC) have emerged. Several of these individual ontologies redefine core concepts of a building already specified in the publicly available ontology version of the ISO standardised Industry Foundation Classes (IFC) schema, thereby violating the W3C best practice rule of minimum redundancy. The voluminous IFC schema with origins in a closed world assumption is likewise violating this rule by redefining concepts about time, location, units etc. already available from other sources, and it is furthermore violating the rule of keeping ontologies simple for easy maintenance. Based on all the available ontologies, we propose a simple Building Topology Ontology (BOT) only covering the core concepts of a building, and three methods for extending this with domain specific ontologies. This approach makes it (1) possible to work with a limited set of core building classes, and (2) extend those as needed towards specific domain ontologies that are in hands of business professionals or domain-specific standardisation bodies, such as the European Telecommunications Standards Institute (ETSI), buildingSMART, the Open Geospatial Consortium (OGC), and so forth.
Full-text available
Background The public health initiatives for obesity prevention are increasingly exploiting the advantages of smart technologies that can register various kinds of data related to physical, physiological, and behavioural conditions. Since individual features and habits vary among people, the design of appropriate intervention strategies for motivating changes in behavioural patterns towards a healthy lifestyle requires the interpretation and integration of collected information, while considering individual profiles in a personalised manner. The ontology-based modelling is recognised as a promising approach in facing the interoperability and integration of heterogeneous information related to characterisation of personal profiles. Results The presented ontology captures individual profiles across several obesity-related knowledge-domains structured into dedicated modules in order to support inference about health condition, physical features, behavioural habits associated with a person, and relevant changes over time. The modularisation strategy is designed to facilitate ontology development, maintenance, and reuse. The domain-specific modules formalised in the Web Ontology Language (OWL) integrate the domain-specific sets of rules formalised in the Semantic Web Rule Language (SWRL). The inference rules follow a modelling pattern designed to support personalised assessment of health condition as age- and gender-specific. The test cases exemplify a personalised assessment of the obesity-related health conditions for the population of teenagers. Conclusion The paper addresses several issues concerning the modelling of normative concepts related to obesity and depicts how the public health concern impacts classification of teenagers according to their phenotypes. The modelling choices regarding the ontology-structure are explained in the context of the modelling goal to integrate multiple knowledge-domains and support reasoning about the individual changes over time. The presented modularisation pattern enhances reusability of the domain-specific modules across various health care domains.
Full-text available
An increasing number of information management and information exchange applications in construction industry is relying on semantic web technologies or tools from the Linked Open Data (LOD) domain to support data interoperability, flexible data exchange, distributed data management and the development of reusable tools. These goals tend to be overlapped with the purposes of the Industry Foundation Classes (IFC), which is a standard for the construction industry defined through an EXPRESS schema. A connecting point between semantic web technologies and the IFC standard would be represented by an agreed Web Ontology Language (OWL) ontology for IFC (termed ifcOWL) that allows to (1) keep on using the well-established IFC standard for representing construction data, (2) exploit the enablers of semantic web technologies in terms of data distribution, extensibility of the data model, querying, and reasoning, and (3) re-use general purpose software implementations for data storage, consistency checking and knowledge inference. Therefore, in this paper we will look into existing efforts in obtaining an ifcOWL ontology from the EXPRESS schemas of IFC and analyse which features would be required in a usable and recommendable ifcOWL ontology. In making this analysis, we present our implementations of an EXPRESS-to-OWL converter and the key features of the resulting ifcOWL ontology.
Full-text available
There is an increasing interest in developing ontological versions of engineering standards. In general, this amounts to restating a given standard in some ontological language like OWL. We observe that without an ontological analysis of the standard, the conversion neither improves the clarity of the standard nor facilitates its coherent application. In this chapter we begin to study the Industry Foundation Classes (IFC), a standard providing an open vendor-independent file format and data model for data interoperability and exchange for Architecture/Engineer-ing/Construction and Facility Management. We first look at IFC and at an existing OWL version of IFC; then we highlight the implicit assumptions and we apply ontological analysis to discuss how to best grasp the type/occurrence distinction in IFC. The goal is to show what has been done in IFC and the contribution of ontological analysis to help increasing the correct understanding of a standard. With this approach, we reach a deeper understanding, which can guide the translation from the original language to OWL with increased conceptual clarity while ensuring both logical coherence and ontological soundness.
Full-text available
Form, function and the relationship between the two serve a crucial role in design. Within architectural design, key aspects of the anticipated function of buildings, or of spatial environments in general, are supposed to be supported by their structural form, i.e., their shape, layout, or connectivity. Whereas the philosophy of form and function is a well-researched topic, the practical relations and dependencies between form and function are only known implicitly by designers and architects. Specifically, the formal modelling of structural forms and resulting artefactual functions within design and design assistance systems remains elusive.In our work, we aim at making these definitions explicit by ontologically modelling respective domain entities, their properties and related constraints. We interpret “structural form” and “artefactual function” by specifying modular ontologies and their interplay for the architectural design domain. A key aspect in our modelling approach is the use of formal conceptual requirements and qualitative spatial calculi as a link between the structural form of a design and the differing functional capabilities that it affords or leads to. We demonstrate how our ontological modelling reflects types of architectural form and function, and how it facilitates the conceptual modelling of requirement constraints in architectural design.
Full-text available
In this paper we propose that adequate treatments of space need to be multiperspectival and related to sound foundational ontologies. To support this, we show that natural spatial descriptions commonly appeal to diverse theories of space and these need to be formally combined to be fully interpreted. Our account draws particularly on the foundational ontology DOLCE and the algebraic specification language CASL. We show how the structuring mechanisms of CASL suggest mechanisms for both building and combining multiperspectival ontologies of space. We also suggest that these mechanisms provide a natural link both with currently emerging cognitive principles such as blending and with developments in ontology mediation and mapping.
Over the past few years, several suggestions have been made of how to convert an EXPRESS schema into an OWL ontology. The conversion from EXPRESS to OWL is of particular use to the architectural design and construction industry, because one of the key data models in this domain, namely the Industry Foundation Classes (IFC), is represented using the EXPRESS information modelling language. These conversion efforts have by now resulted in a recommended ifcOWL ontology that stays semantically close to the EXPRESS schema. Two major improvements could be made in addition to this ifcOWL basis. First, the ontology could be split into diverse modules, making it easier to use subsets of the entire ontology. Second, geometric aggregated data (e.g. lists of coordinates) could be serialised into alternative, less complex semantic structures. The purpose of both improvements is to make ifcOWL data smaller in size and complexity. In this article, we focus entirely on the second topic, namely the optimization of geometric data in the semantic representation. We outline and discuss the diverse available options in optimizing the data representations used. We quantify the impact of these measures on the ifcOWL ontology and instance model size. We conclude with an explicit recommendation and give an indication of how this recommendation might be implemented in combination with the already available ifcOWL ontology.
It is widely accepted that modularity will help to solve many problems in the construction and management of ontological systems, and researchers are actively investigating techniques for modularization and parameters for module assessment. This note aims to diverge attention from technical issues to what is the basic goal of modularity. Modularization is a technique that can be devoted to solve different types of problems, depending on the type of ontology one works with. We start from a classification of ontology research in three rough classes, and discuss what modularization should achieve from each perspective. This observation gives already some indications on what modules one should look for. We conclude with the optimistic view that the modular approach can radically change the way we build foundational ontologies. This, however, requires a further study of what should count as a module. Something we are still far from understanding.