See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/305953953
Ten questions concerning building information modelling
Article in Building and Environment · August 2016
DOI: 10.1016/j.buildenv.2016.08.001
CITATIONS
113
READS
8,470
1 author:
Žiga Turk
University of Ljubljana 193PUBLICATIONS 2,233CITATIONS
SEE PROFILE
All content following this page was uploaded by Žiga Turk on 03 May 2018.
The user has requested enhancement of the downloaded file.
1
Turk, Žiga (2016). Ten Questions Concerning Building Information Modelling, Building and Environment, Elsevier, doi:10.1016/j.buildenv.2016.08.001
Ten Questions Concerning Building Information Modelling
Žiga Turk,
University of Ljubljana, Faculty of Civil and Geodetic Engineering, Jamova 2, 1000 Ljubljana, Slovenia [email protected]
Abstract: Building information modelling (BIM) has been a dominant topic in information technology in construction research since this memorable acronym replaced the boring “product modelling in construction” and the academic “conceptual modelling of buildings”. The ideal of having a complete, coherent, true digital representation of buildings has become a goal of scientific research, software development and industrial application. In this paper, the author asks and answers ten key questions about BIM, including what it is, how it will develop, how real are the promises and fears of BIM and what is its impact. The arguments in the answers are based on an understanding of BIM that
considers BIM in the frame of structure-function-behavior paradigm. As a structure, BIM is a database with many remaining database challenges. The function of BIM is building information management.
Building information was managed before the invention of digital computers and is managed today with computers. The goal is efficient support of business processes, such as with database-
management systems. BIM behaves as a socio-technical system; it changes institutions, businesses, business models, education, workplaces and careers and is also changed by the environment in which it operates. Game theory and institutional theory provide a good framework to study its adoption.
The most important contribution of BIM is not that it is a tool of automation or integration but a tool of further specialization. Specialization is a key to the division of labor, which results in using more knowledge, in higher productivity and in greater creativity.
Keywords: building information modelling, BIM, building information management, automation in construction, computer integrated construction.
1 Introduction
Humans do many things without planning or thinking, such as when someone throws us a pen we asked for, we catch it without consciously calculating the curve-of-flight and planning the catch. However, when humans do something rationally, we first imagine doing it. We act it out in our heads before actually taking action. When hanging a painting on a wall, for example, we imagine where it should hang, how it should be aligned with existing paintings, where we need to drive the nail into the wall, considering the
location of the hanger and the offset to the edge of the frame.
As Aristotle put it [1] “First, have a definite, clear practical ideal; a goal, an objective. Second, have the necessary means to achieve your ends;
wisdom, money, materials, and methods. Third, adjust all your means to that end.” In a similar fashion, Shakespeare [2] described the process of construction: “When we mean to build, We first survey the plot, then draw the model; And when we see the figure of the house, Then must we rate the cost of the erection; Which if we find
2 outweighs ability, What do we then but draw anew the model, In fewer offices, or at last desist To build at all?”
Information models of buildings have been represented as drawings since paper became inexpensive and available for tasks such as building design. As the Shakespeare citation demonstrates, drawing was called modelling well before modelling is replacing drawing.
Paper enabled fairly reliable communication among the designers of buildings and bridged distances in time, space and profession [3].
The latter is particularly important because it enabled specialization of professions. Earlier, master builders—with all the relevant general knowledge and all the specific plans – were, due to the lack of communication, mostly confined to themselves. Later they could be replaced by teams of specialists. Teams could collaborate because paper-based communication offered a reliable way to communicate [4]. They were able to share an information model of a building. On one hand, this enabled the specialization in the professions designing, constructing and
managing buildings. It enabled the specialization of businesses involved with these processes.
However, it also resulted in fragmentation and disintegration of professions, knowledge, processes and businesses. The process accelerated with the introduction of digital technology. A research topic called computer- integrated construction [5,6,7] set out to develop solutions that would counter the fragmentational effect of digital technology.
The main topic in computer-integrated
construction research was the development of methods to describe buildings using a common language and methods to collaborate in that language. In the late 1980s, the solution was called “Conceptual Modelling of Buildings” [8].
Later, the community borrowed the term
“Product Modelling” [9], which was used in mechanical engineering to describe design and manufacturing information about future
products [10,11,12]. The ISO-STEP standard first appeared in mechanical engineering [13]. There
were several attempts to use it in construction [14,15] but without much direct practical impact on the industry.
In the meantime, the construction software industry was creating various tools to support engineering activities during designing and planning. Most of that software was concerned with the analysis and simulation required by various specialists. They operated on
mechanical, physical or mathematical models of phenomena, such as finite element models of beams, energy models of walls, and process models of work. These models became
increasingly accurate, and the simulations, due to the increasing speed of computers, became more reliable.
Other software was concerned with replacing paper-and-ink drawing boards, and drawing software evolved. The key evolution was in the elements that the drafter could place (draw) on the canvas (drawing). The simplest drawing elements are pixels, followed by lines and other 2D geometric shapes, followed by 3D geometric objects. In some fields—for example for
organizational diagrams, software engineering and industrial process maps—2D geometry was replaced by symbols that stand for something in the problem domain – such as sectors and their bosses, machines on the assembly lines, their inputs and outputs, or steps in a computer algorithm.
Finally, 3D CAD software evolved from allowing the modelers to place 3D geometric objects into 3D model space, to placing engineering or architectural elements into the digital
representation of the landscape. In a geometric CAD system, it was the human who had to interpret, for example, a cylinder as a structural column. In BIM software, this is explicitly stated in the resulting database. The software industry began to transition from CAD to BIM.
While the acronym BIM is attributed to Jerry Lassarin [16], the concept was a result of a long series of research under the topic of conceptual modelling of buildings and product modelling of buildings since the 1970s [17,18].
3 Conceptual modelling, product modelling and BIM have traditionally been subject more to research push than industry pull [19]. The first attempts to standardize data structures needed to describe the built environment came from the top-down within ISO-STEP, followed by a more bottom-up approach in the International Alliance of Interoperability (IAI) with Industry Foundation Classes (IFC) [20,21].
In summary, designers of buildings have always used information models of buildings. In fact, the design process was all about information modelling of buildings. With information
technology, the information models first became digital and have since become increasingly well structured. CAD evolved naturally from 2D geometry via 3D geometry towards 3D professional objects with the 4th dimension added for time. The amount of data that we have on buildings is growing exponentially, much like Moore’s law and IT capacity allow.
Specialized engineering software (such as finite elements software) was based on engineering objects—not lines or pixels—from the very beginning.
2 Ten questions
BIM is an exhaustive research topic in the field of construction informatics [22] or computing in building engineering. Selecting ten questions is not easy. The author chose questions that may generally be overlooked in the ongoing quest for
“better” models and more “efficient”
collaboration. The questions can be grouped into three categories:
1. What BIM is and what are its future directions?
2. What are the fears and promises related to it?
3. How it is affecting selected areas related to construction?
The author is aware that each of these questions deserves a separate in-depth study and hopes they would follow and complement the problem solving type of research that dominates the BIM research community.
2.1 What is BIM?
The acronym stands both for building
information modelling (the process) and building information model (the artefact), and the attention of the research community and software developers alternates between the two. Initially, the challenge was the
representation of buildings. As the
representation matured, the attention shifted towards the processes in which these
representations can be created, developed and used.
BIM can also stand for building information management—the control of the processes in which models are built and used [23]—and for building information marketing [24]. The latter is a cynical observation of the exploitation of the acronym, both in the industry and in the academia..
BIM initially represented the new, structured ways to represent buildings that went beyond lines and “implicit meaning” towards objects and
“explicit meaning”. The issue of meaning will be revisited in Section 2.6.
One could argue that building information modelling—the process of creating information about (future) buildings—has existed for centuries. This is even acknowledged in the BIM maturity levels [25,26], where BIM level 0 corresponds to information modelling of
buildings using paper drawings, BIM level 1 to 2D and 3D CAD and only BIM level 2 and above to object-oriented (see Section 2.2.1)
representations of buildings and corresponding processes.
The US National Building Information Modeling Standard defines BIM as “A digital
representation of physical and functional characteristics of a facility… and a shared knowledge resource for information about a facility forming a reliable basis for decisions during its life-cycle; defined as existing from earliest conception to demolition”[27]. In this definition, the M clearly stands for model. The definition is broad, and many legacy
4 technologies and techniques could fit into the definition.
Some define BIM in broader terms—as an approach to engineering collaboration rather than specific technical solutions [28]: “Building information modeling (BIM) refers to a
combination or a set of technologies and organizational solutions that are expected to increase inter-organizational and disciplinary collaboration in the construction industry and to improve the productivity and quality of the design, construction, and maintenance of buildings.”
The two ways of defining BIM can be reconciled using an approach engineers and architects have been using for a long time. When describing things, they often resort to a paradigm known as structure-function-behavior [29,30]. Using this paradigm, the definition of BIM has three facets:
structural, functional and behavioral.
1. Structural defines how it is organized, what parts it has, and how they work together.
2. Functional describes how it can be made useful.
3. Behavior describes how it responds to its environment.
2.1.1 Structure of BIM
Structurally, BIM is a structured representation of a building, for example, an object-oriented representation of a building. The phrase “object- oriented” should be understood as in object- oriented analysis, design and programming. In the object-oriented paradigm, the problem domain is seen as a number of objects. The objects are representations of real world items (1) that have an identity, (2) that we know something about and (3) that we can do something to. They may belong to classes and include or relate to other objects.
In BIM, we know more about real world objects than their geometry. The more we know the better. We should be able to do more to the objects then move them around. For example, we should be able to put a load on them and see
what happens or let the sun shine on a building and see how it heats up. This is because objects include behavior.
Ideally, the structure of BIM would be that of an object-oriented database. The requirements of a database in general include [31] sharing data in multiuser system, support for multiple views of data, controlled data redundancy, enforced integrity, restricted unauthorized access, data independence, transaction processing, backup and recovery. Interestingly, this is in fact a superset of early requirements for product models of buildings [32].
2.1.2 Function of BIM
Functionally, BIM is a communication backbone and shared source and destination of
information required by and created by the individuals and processes that participate in building processes that enables “gains in saving in cost and time, much greater accuracy in estimation, and the avoidance of error, alterations and rework due to information loss [33]. Ideally, all information about a building during its entire life cycle would be created in a BIM process and stored in a BIM database. The function of BIM in engineering and architectural processes is the same as the function of
management information systems (MIS) for management processes. Management information systems have been traditionally implemented with database management systems (DBMS). In this sense, the function of BIM is building information management.
2.1.3 Behavior of BIM
The behavior of BIM is that of a socio-technical system—a group of interacting, interrelated, or interdependent elements or parts that function together as a whole to accomplish a goal, that has a boundary and, through inputs and outputs, interacts with the environment [34,35]. The parts were addressed in the structural definition, the goal in the functional. The interaction with the environment is addressed in this subsection.
There are two contexts in which BIM exhibits behavior: the context of a specific project and the broader general industrial and societal
5 context. In a project context, BIM pushes
technological opportunities and changes business models, organizational patterns and information processes in which the project is developed. It encourages centralized
information management, reduces
improvisation, and calls for better organization of processes. It should respond to varying levels of IT literacy of project partners and their achieved technological levels; it should be a means to an end rather than a goal in itself.
On a general level, BIM encourages changes in legal processes (BIM as a representation of designs when interacting with authorities) and changes in procurement processes (BIM mandated by the investors) and is causing restructuring of the industry (vertical design- build integration so that the benefits of BIM in the construction phase are reaped by the same company that made a greater effort due to BIM in the design stage). The development of BIM systems responds to business needs, value proposition [36,37] and the demand for return on investment [38].
2.2 What are the future directions of BIM?
The future of BIM is determined by the ongoing research and software development that has been summarized in several studies [39,40,41].
The contribution of this section is to look at these developments through the lens of the structure-function-behavior of BIM.
2.2.1 Structure – how to make the database better
The evolution of structure is in the development of the database, which includes four directions:
1. Better database schema.
2. Better database features.
3. Progress in database coverage - lesser use of information storage outside of a database.
4. Development beyond the object- oriented data model.
Progress in objects and schema. Structurally, the progress is in objects that are richer (more
attributes and functions) and more complex (more relations to other objects). This process will never be complete. It is impossible to predefine a data structure (define a schema) that can include, in a structured format, all information about buildings that at someone would like to have. There are volumes of philosophy that argue that this is not possible [42]; therefore, this process will never be complete. On the other hand, it is likely that in the near future, the schema will be sufficient for a particular type of building, such as temporary buildings made of containers, pneumatic buildings, prefabricated homes of a certain manufacturer, family houses of a certain type, motels of a motel chain that look the same, and shopping malls.
Improving the database features of BIM, as standard in databases proper.
Progress is being made in the sharing of data in a multiuser system as in a true multiuse database with several users simultaneously changing the database, such as in management information systems. This is approached through BIM servers [43]. The underlying database has to give up proprietary file containers and rely on capable, industry-standard object-oriented, relational or RDF databases.
Support for multiple views of data is approached through model views [44,45], which are a well-established concept of BIM used to extract data for certain disciplines or use cases. In the traditional database parlance, a view is usually a concept associated with the retrieval and presentation of
information; in BIM, views are also used as the context in which information is modified.
Controlling data redundancy is related to the desired feature of BIM that each piece of information is stored only once.
The problem has various levels of complexity, from a simple desire that the width of a door is only stored in one
6 location in the database, to making sure that all doors, if they are the same, are stored only once, to parametric links among attributes (a door is as wide as a window) as well as to links to functional requirements.
Restricted unauthorized access means providing access control on various levels of granularity for who can do what. Again, we take this for granted in proper databases. Scenarios in
engineering and architectural design are more complicated, but the principle is the same.
Data independence is a concept that shields applications through which users change the contents of a database (e.g., Revit, ArchiCAD) from changes in the schema of a database (on a BIM server), the ways it is organized, and where and how is it stored. In model-view-
controller terms [46], the BIM community needs to investigate the decoupling of the “controller” (Revit, ArchiCAD) from a “model” (BIM server).
It will also need to address the problem of opening a model created with an old version of a modelling tool with a new version of the modelling tool. If data independence is achieved, this should not matter.
Transaction processing is a database concept that organizes interactions within a database into unbreakable, complete groups. In BIM, this approach is needed when several actors are working on the same parts of a building at the same time. Current approaches to locking parts of the model or checking worksets in and out perform this task on a rudimentary scale. Proper transaction processing simplifies backup, recovery, and going back in time, which is a feature that should replace the creation of daily copies of the model.
Progress in database coverage. The third direction is in the reduction of stores that are
not in a database, e.g., information about a building that is not in the model file. To what extent this is desirable will be discussed later. In database parlance, this too would be about reducing redundancy.
Progress beyond object-oriented models.
Object-oriented structuring of information is not the only way to represent building information.
Large swaths of information can be represented using relational models (e.g., COBIE). In the future, computers will continue to become faster. This will allow for representations that are less efficient than object-oriented models and that can represent more complex
relationships, such as RDF, OWL, predicate logic and other approaches from artificial intelligence.
Research on this topic exists but is not practical for large realistic projects at the current stage of technology [47,48].
2.2.2 Function – how to use the database better
Functionally, BIM has the ambition to expand the coverage of the building process both in depth and in breadth. BIM started as a detailed design-stage technology. Both research and commercial tools now use information models from inception to the maintenance of buildings.
The depth will increase in a way in which an increasing number of vertical applications—
specialized analysis and simulation tools—will establish two-way communication with the information model. Standards will continue to play a role here but are not vital. We will revisit the issue of standards in Section 2.4.
From a structural and functional perspective BIM will approach the impossible—a centralized shared database to represent a building. It will not be possible to achieve all the attributes of a good database as described in Section 2.1.
Functionally, it is impossible and impractical for all building processes to have their information needs satisfied through such a database.
A likely scenario, like with the structure, is that BIM will be sufficient to cover practically all tasks related to a finite, limited type of building and
7 that software vendors will offer an integrated suite of applications that will perform the necessary simulations and analysis for that one, well-defined type of building.
2.2.3 Behavior – how to change the environment and be changed by it The industrial environment will shift the behavior of BIM towards what is considered practical. Some of those practicalities are addressed in the following sections.
2.3 Will it ever be possible to describe all that has to be known about buildings in BIM?
Based on the answers to previous the questions, particularly on the topic of the structural
evolution of BIM, the answer is yes!
However, this will be achieved not because of a belief in technology but because of a belief in engineers, architects and builders. It was
possible to describe all that anyone had to know about a building using paper drawings. The proof that it was all they needed to know is in the existence of buildings, almost all of which have successfully been built without the benefit of BIM. It is an essential feature of engineers (as opposed to some other professions) that we can address incomplete information. BIM will remain incomplete, and this will not be a problem.
The answer is also no. No, it is not possible to think in advance of every possible question a person involved with the built environment might have and expect to find the answer inside a BIM database. Impossible goals require infinite effort, both in research and in practice. While researchers may find impossible effort desirable because it sustains research opportunities, practice will be looking for practical solutions requiring limited effort.
The lesson of this is that there is a limit to how useful the increasingly complex BIM models are.
The quest for an “ever more complete” model may be the wrong quest. A likely and still useful scenario is that building information will continue to be scattered over a few databases and several files. For simple, routine buildings, a
single database scenario is easier than for unusual, complex buildings.
2.4 How important are BIM standards?
The idealistic holy grail of BIM has been to create a common language for builders. Not quite unlike the common language that was lost in biblical times when God mixed up the
languages of those that were building the Tower of Babel. Historically, many development concepts related to BIM occurred in the context of standardization. The idea in the 1990s and early 2000s was that the standard for the description of buildings should be defined, and the industry should write software that
conforms to it – both in the information exchange and as a basis for the internal representation of buildings in software. Some essential theoretical and developmental work has been performed in this context [49,50]. This is known as top-down standardization: first something is defined and then it is commercially implemented.
Much of the IT software industry, however, has been using bottom-up standardization. There were several candidates for the language for the Web, but sometime in the early 1990s, HTML emerged as the de-facto standard. In a similar manner, PDF and XDOC emerged as standards for documents and MPEG for music and video.
Software vendors have created their own ways of representing buildings. Some allowed for better or worse import and export into the IFC standard. However, in principle, they were aware that they were not competing just on how fast their modelling engine works, how realistic the renders are, how useful the schedules are, and how good their IFC export is. Software developers are in a competition for which tool can produce better, richer, and smarter information models. Their own schema is a key element in this competition, and they invested significant effort to create one. They are also in a competition, who can provide a friendlier environment for other software developers to
8 write software that would be interoperable with theirs.
2.4.1 Not very important
This effort of the software industry is not a bad thing. Proprietary schema are, as a rule, richer.
They are not a compromise. They are not theoretical; they are practically implemented in the next version of the software. And not much harm is being done. The future direction of interoperability will not go in the direction of every possible engineering software working with every other software exchanging data in IFC or some other standard. It is theoretically impossible for these exchanges to work seamlessly both ways [51] if the schema are different. Schema will be different. BIM models are not as simple as JPEG files where thousands of programs can open and modify them.
A more likely scenario, which is unfolding, will be that of illustration and office software. Decades ago, there was different software for databases, spreadsheets, documents, and slides. This software had different vendors. It was not a committee that resolved the problem of
inserting a spreadsheet into a document but the fact that Microsoft, Apple and Open Office created their own suites of software that worked well within each other’s ecosystem. We are observing a similar process with engineering software that is now grouped around the models of major vendors and accessing the vendor- specific model through vendor-specific application interfaces (APIs).
Another breakthrough that is making BIM standards less important are technologies around XML. Earlier it was indeed very hard to write (n-1)2 translators among n different programs. A neutral, standard schema and format looked like an elegant solution. However, in a realistic scenario there is no need for each software to talk to all others. There is little need, for example, for a structural engineering
simulation software to exchange information with building physics simulation software. Both just need to talk to the model server.
Secondly, XML technology with related open source libraries makes the task of writing translators significantly easier. XML technology trivializes the interoperability on the syntactic level and, with the ability to encode and process database structure and semantics, it significantly eases the mapping of the data structures from on to another application without a detour via a standard schema.
2.4.2 Quite important
There are three types of standards that are important: (1) standards for the definition of concepts (not standard definitions of building specific concepts), (2) standards for the definition of syntax and (3) standards for the definitions of APIs. These are generic standards used in service-oriented architectures across industries. Because writing interfaces, parsers and translators became much easier over the last decade, standards for AEC concepts themselves, i.e., the building elements, are not as important as thought in the beginning. For interoperability, proprietary but open is not much worse that standardized and open. The proof is in the existence of many interoperable programs around popular modelers that exchange information using native file formats or (more likely) native APIs. For example, over 500 such programs and apps are listed in the Autodesk APP Store and work with Revit [52].
Based on the experiences with service oriented architectures that ensure the interoperability of various internet services, the schema of a building model, or at least the application interface to a building model, should have the following features (listed in the order of importance):
Open. Others should be able to learn the details of the API and schema. It should not be a secret.
Machine readable. A machine-readable schema definition allows software generators to create interfaces to databases implementing the schema.
Use standard data and schema
representation language (like XML, XML
9 schema, RDF, OWL) so that generic tools can be used.
Be compatible with a standard schema (like IFC).
However, even in the scenario presented, there is an important role for standards such as IFC and OpenBIM:
Standards provide a reference or starting point for the software developers, i.e., a schema to be improved upon.
Standards can provide the lowest common denominator for information exchange among software that did not choose to interface with the proprietary schema or proprietary API. In BIM, they can play the role of DXF that has been the lowest common denominator for exchange of CAD information in architecture and engineering.
Standards provide a neutral representation that authorities can demand for procurement and permit processes. It is not possible for authorities to ask for information in a format that is proprietary.
Standard formats are safer for long-term preservation of information. It is much more likely that information in a standard format will be readable after decades or even centuries than
information in the format of a software vendor that happened to be market leader at the time when the building was designed.
Standards provide an environment for the publicly funded academia to contribute to the progress of BIM technology in a vendor neutral way.
The above justifies the continuing development of the standards but the focus is less on
interoperability.
2.5 Will the result of BIM be computer integrated construction?
One of the often-cited problems of the construction industry is its fragmentation. It is apparent both in information-related tasks and during the construction phase. Several
companies, large and small, are involved, including many specialized consultants. The construction industry is lacking well-designed, managed and stable supply and delivery chains known in the manufacturing industries.
It has been a long-standing hope that through information technology, the construction industry will become more integrated. This hope could be wrong. Way back in history, when almost no information technology was used, almost no documentation existed and all decisions were made at the construction site, construction was quite integrated. The material world and material processes were driving the integration—people had to come together to work on the same physical object.
The fragmentation of the industry started with the introduction of information technology, such as paper drawings because information
technology allowed for specialization and specialization allowed for more and different knowledge to be incorporated in the processes.
Specialization accelerated with digital information technology and so did the
fragmentation by any objective criteria, such as the number of specialists or number of
businesses involved. It was the physical building that was holding it all together.
Increased fragmentation brought in more specialized knowledge, and the greater
knowledge resulted in better quality and higher efficiency. Adam Smith, in Wealth of Nations [53], noted that specialization through division of labor is “the cause and source of prosperity”.
In this context, the goal of BIM is not computer- integrated construction. The goal of BIM is to continue the centuries’ long trend of
information technology that allows for
10 increasing specialization and division of labor.
Information technology makes collaboration of more people from more specialized professions possible. Because the real (material) structure being built forces the integration of material processes, with BIM, the industry obtained a core element that holds the information processes together. The stronger the pull towards a digital model, the more professionals, software and tools can be deployed to make the designs and plans better and faster.
Information technology, therefore works both ways. It enables greater fragmentation while holding the processes together to contribute efficiently to a single goal.
2.6 Is there such a thing as semantic BIM?
There are four qualitative levels of
interoperability: technical, syntactic, structural and semantic [54,55].
Technical means that there is a data
transportation service that can carry information from A to B.
Syntactic interoperability means that the syntax of information to be exchanged or shared is defined and programmers can write code that will extract the necessary pieces of data. To put it simply, one application is able to read another application’s data, which makes them
interoperable. Being able to read DXF files is an example of this type of interoperability.
Structural interoperability means that the data structures of two applications are compatible—
the schema used is shared or at least understood. An example is the IFC standard.
Semantic interoperability implies a shared understanding of information so that the
“meaning” of information in two interoperable applications is the same. Semantic
interoperability is considered the next step in the evolution of interoperable building software [56]. The goal is to “make building information models understandable and model data sharable
across multiple design disciplines and heterogeneous computer systems” [57].
2.6.1 Meaningful to machines
Semantics is a heavily used term in computer science. Its popularity received a boost with the concept of the semantic web in the beginning of the century. In philosophy, semantics is the study of meaning. Using the term as an adjective, such as semantic web, semantic interoperability, and semantic BIM, implies that the bits and bytes of web pages, exchanged files or building information models would have meaning. The key issue here is to whom this information will be meaningful—to software or to humans.
The promise of semantic X is to make X
meaningful to computers. Semantic BIM would be BIM that would have meaning for computers or computer software. This will be possible only after computer software gets a much broader sense of the world than they have today and would be capable of finding out that an object actually represents a column without being called so.
To humans, lines on paper have meaning, as do objects in an information model. The lines representing a wall or an object with the name ifcWall have meaning because the symbols become associated (the first in its graphical representation, the second because of the word
“wall” used in the name) with a real world object with which we have real world experiences: such as living within walls, painting walls, crashing into walls, and building walls.
Humans learn the meaning of things because they can relate the three apexes of the meaning (semiotic) triangle [58]: the symbol (ifcWall), the real world object and the ideas about walls in our minds. Software only knows about symbols and is confined to one apex of the triangle.
Closer to connecting at least to apexes of the meaning triangle are technologies that include remote sensing or laser scanned point clouds where software is able to recognize objects from sensor data [59,60].
11 The goal of semantic interoperability is to agree on the “meaning” of symbols. Claiming that software learns the “meaning” of objects or that the meaning is encoded in language structures such as RDF and OWL is a misuse of the term
“meaning”. The graphs of OWL or RDF look meaningful to humans because the arches and nodes have natural language labels, such as
“has-part”, “has-property”, and “rdf:type”.
We could talk about semantic representations of buildings if software was able to determine that an object that has three real numbers as attributes is a point in space without directly providing the information that the object is from a class called “ifcPoint”. This is difficult for points and is levels of magnitude more difficult for complex engineering objects.
2.6.2 Meaningful to humans
A lesser goal of semantic interoperability is the shared meaning of symbols among humans.
However, this too may be next to impossible. As Haushofer and Neuhold note, [61] “As long as people are the designers of models there will always exist different conceptions and interpretations, even for superficially
homogenous domains and application contexts.
We therefore believe that computer science research should take this situation into account and find solutions that deal with a multitude of models and allow for their reconciliation. The establishment of mappings between existing models is such an approach”.
Therefore it is not entirely substantiated to call a BIM semantic BIM. The meaning emerges in humans because they can establish relations between symbols that computers display, the real world objects and the concepts in their minds [62]. Interoperability among software can be achieved on the level of technical
connectivity, syntax and structure. This is not to say that software relying on building information models cannot become increasingly smarter and incorporate increasing amounts of engineering and architecture information. This will be particularly important in vertical applications
related to structural systems, building envelopes, and sustainability.
2.7 How is BIM affecting building processes?
In the past, construction was defined through two sets of documents: the design documents that described how things should be and the planning documents that defined who should do what and when. The two sets of documents were a result of information processes—
designing and planning. This section focuses on how the information processes are changing because of BIM.
2.7.1 It makes them more rigid
Building laws defined, on a page or two, what documents were needed in what phase and what they should contain. The building process was described quite briefly by laws or by professional associations, such as the Royal Institute of British Architects (RIBA). RIBA has published the Plan of Works since 1963 [63]. The first plan was on one sheet of paper and
described roles of participants in design in construction and the information exchange among them. At the time, the exchanges were documents. The unit of coordination and planning was a document. Such an “exchange”
was a document that would be submitted to the building office for approval. The processes were defined rather broadly [64,65]. Within each process, much just-in-time improvisation occurred [66]. Rigid schema applied only to major stages in the process, such as appraisal, design brief, concept, design development, technical design, product information, and tendering.
In a BIM environment, thousands and thousands of digital objects are created and refined. Access to these objects is shared, and they are created in a collaborative manner. The processing paradigm—one in which work is modelled as a process with inputs, outputs, controls and methods—is used to organize activities which are much finer grained.
12 Inflation of documents, such as BIM protocols, BIM execution plans, master information delivery plans, BIM coordination programs, project information plans, and asset information plans, that define data exchanges, data drops, access rights, levels of detail, levels of
development, grades of development and depth of detail, illustrates the increasing organizational and managerial complexity, which is a direct consequence of BIM. Counting purely how much detail of the processes is defined supports the thesis that processes are more rigid than before.
It would be interesting to objectively compare the complexity metrics [67] of the processes before BIM and after BIM.
The phenomenon that information systems reduce the flexibility of the processes that they support was identified already in the 1980s, when a broad adoption of management
information systems was occurring. Similar levels of automation and IT support are now being achieved in the building processes. They reduce the flexibility, need and opportunity for
improvisation which could be efficient given the fact that buildings are unique.
2.7.2 It should change our understanding of activities
Perhaps the problem is in the reliance on the understanding of human activities as processing.
A processing paradigm with arrows connecting the processes is well-suited for exchange
collaboration architectures where information is exchanged, such as by sending a drawing or a file from A to B. Sharing is different. A process is not initiated with the receipt of information. An alternative paradigm looks at building (verb) as a socio-technical activity and captures social interactions.
The term “information” in the phrase “building information models” should be reconsidered in its pre-technological definition, which states
“information is informing the reader”. In the past, with less computerized documentation, information was exchanged to inform co- workers and to make them do something. The difficulty of exchange limited the amount of
information exchanged to what was essential.
The internet eased the transfer of information, information modelling increased the structural complexity of the information, and cloud infrastructure eased the sharing of information.
Everything is available to anyone at any time, not just what is needed “to inform the reader”.
If an architect sends a structural engineer a conceptual architectural design, it is not just to inform her, it is also a speech-act saying “here’s my design, does it work for you?” Updating the architectural objects in a BIM database carries no such pragmatics. The “need to know” has been replaced by the “need to share”. And this is not always a good thing.
2.8 How is BIM affecting education in the field?
The question goes well beyond teaching BIM [68,69], which is a must both in regular as well as in life-long education in AEC (architecture, engineering, construction). The interpretation of BIM that sees it as an approach to building and as the behavior of those involved in the building processes (Section 2.1.3) provides an argument for a substantial overhaul of education in AEC.
However, not all topics are affected equally.
Therefore, the short answer to the question is “a lot”. The more detailed answer addresses which areas and topics are affected and how.
2.8.1 Computing
This includes the courses in which engineers and architects learn about information and
communication technology, computer science, and design communication. In these courses, students learn about the principles of computer- integrated construction and the related tools and work methods. Many fundamentals and theories remain the same. More emphasis on theory should be given to information modelling, conceptualization, databases and object
orientation. However, the shift from tools for generic objects (CAD drawings, databases) to tools that address industry-specific objects requires a project-based approach. This includes linking IT courses with professional courses about buildings, building envelopes, structural
13 elements, and materials. Virtual studios and
project-based learning [70] are required and are enabled by BIM.
Computer programming should be gaining importance because BIM uses objects and objects are not just about data, but about methods. i.e., programs. Essentially, there is no difference between specifying where an object should be placed or writing a method that will compute, in real time, where it needs to be.
Being able to program that, as some software allows, is important.
The third body of knowledge in this area is education for BIM-related professions: BIM modelers, BIM operators, BIM coordinators and BIM managers.
2.8.2 Project management, organization BIM is fundamentally changing the organization of building processes (Sections 2.7 and 2.9), workflow, contractual arrangements and legal relations [71]. Courses on construction collaboration, construction management, construction law, and cost estimating should prepare students for this new reality in the industry.
2.8.3 Designing
Designing buildings is a core activity of engineers and architects. A potential victim of BIM could be design thinking [72], which is [ibid.] “a methodology … where innovation is powered by a thorough understanding, through direct observation, of what people want and need in their lives and what they like or dislike …”. It is a way in which designers think when designing things—from bottles to skyscrapers. Designing can involve various methods and approaches to reasoning and logic, but in nearly all cases, there is a separation between form, function and behavior. As an early reference model for building information models put it, there is a distinction between functional units and technical solutions [73].
The danger of BIM and of education using BIM is ignorance of the part of design thinking that is involved with function. BIM software makes it
too easy to place elements into the model (walls, columns, ceilings) that are technical solutions (not functional units) and structures (not functions). This issue is not solved with complex schemas for levels of development (LODs) because even the simplest and most generic LOD states “there is a thing”. A thing is a structure, it is not a function. Another way of putting it is that BIM is answer driven not question driven.
“The fear is that heavy emphasis on "how to"
guarantees a loss of the critical "why”” [74].
2.8.4 All other courses
BIM defines the concepts that we use to describe buildings. BIM software defines the vocabulary of the elements out of which
buildings are composed. Conceptual information models underlying BIM software define objects and attributes in a very precise, formal way.
They define what is a structural wall or column and what information we have about it.
Software for modelling and for analysis increasingly uses the same language. A similar transition should occur with the language used in professional engineering courses, which teach about the same objects.
These concepts could be mapped to concepts that are taught in undergraduate and graduate courses. These concepts will appear in the software in which that knowledge is embedded.
BIM is changing the language of construction and therefore impacts the language of
education. The change, however, goes beyond language. BIM objects will be increasingly smart;
they will not be information but knowledge objects embedding professional knowledge.
BIM, when understood in an object-oriented way, does not just define the vocabulary, objects and their properties. It also defines the methods that the objects can perform. Some of the functionality of the methods is incorporated in the models implemented in BIM modelers. Some other functionality is “outside” of the model, in applications such as BIM model checkers (BMC) [75]. The best known is Solibri [76]. Software like this is increasingly incorporating the engineering knowledge that students of engineering and
14 architecture are learning. It can check designs against building codes or codes of practice, in addition to analyzing the purely geometric correctness of the model. This too needs to have an impact on the pedagogy, for example, to restructure knowledge in a way that can be incorporated into model-checking software.
A consequence of all this will be the increasing automation of simple and routine engineering tasks. A deeper understanding of architecture and engineering, as well as the environmental and societal impacts, will remain in the human domain.
2.9 Why is the industry not adopting BIM faster?
This is a question as old as the first working prototypes of BIM, which emerged in the 1990s—scientists wondering why the industry does not use this superior technology [77].
Among the answers, a lack of software
applications, a lack of trained BIM personnel and a lack of a legal and insurance framework that supports collaborative project delivery are cited [78]. Others cite the conservative nature of the industry and the entrenchment of drawings [79].
Several authors have proposed that BIM should be advertised through education or mandated by governments. However, if BIM is as good as advertised, why should its use in private
investments be mandated by law? There was no government action to force businesses to start using the Web, mobile phones, and
management information systems. Some businesses did and succeeded; others did not, failed and lost in a competitive market. Industry is rational; it uses what is useful. The use of BIM is increasing [80,81,82] as technology matures.
Additionally governments are demanding the use of BIM in their procurement to save taxpayers’ money.
In addition to considering the lack of tools and knowledge and cultural reasons, the problem should be studied from the perspective of economics and organizational science.
The economic response would be in the restructuring of the value chains so that the party investing in quality BIM models also reaps the benefits. The added effort is in the design studio, but the benefits are reaped at the construction site with less unexpected work, fewer collisions of building elements and workers, and fewer requests for information.
One way to approach this goal is in the vertical integration of the industry, as in the design- build-(operate) business model [83]. There are reports that in some countries, the restructuring of the industry in this direction coincided with the broad introduction of BIM [84].
2.9.1 Game theory and institutional theory
The proper theoretical tool to study the economic factors that encourage or discourage the use of BIM is game theory [85,86], which started as a study of phenomena where an optimum is not achieved if each party pursues just their own best interests. This is exactly the scenario if BIM is operated in an industrial setting where the players are neither vertically integrated nor coordinated by a game
coordinator who could introduce agreement on how the game should be played. In a BIM setting, the game is a BIM process and the game coordinator is the BIM coordinator.
The other overlooked approach to study BIM adoption is through the eye of theories that study the change of institutions: institutional theory [87]. The term institution in this context means any type of organization that people organize, including construction projects, construction companies and government authorities for the approval of building designs [88]. This theory studies how institutions change because of technology and the hurdles and obstacles due to the entrenched ways of doing things that are present in the institution’s DNA.
These studies explain many of the phenomena in BIM uptake [89,] but more can be done in this area.
15
2.10 Does BIM boost or harm creativity?
When I was urging a young student of architecture to use a BIM modeler instead of drawing software she refused with disgust. “It is just like designing houses in the Sims computer games for youngsters” [90]. A few years later, she is now using BIM, mostly because it makes her more productive. Before BIM, her designs were influenced by how easy they were to cut from wood or foam to make physical models.
BIM software allows her to produce designs that are easy to make using BIM modelers. This anecdote illustrates several concerns and hopes related to BIM and creativity.
But what is creativity? In theory, designing can be routine, innovative and creative [91]. The meaning of these three terms can be defined in at least three different contexts:
in the context of the design process
in the context of the design language
in the context of the design results 2.10.1 Design process
A traditional view of the design process is that it is a search in the space of potential solutions. In routine design, the search is constrained to a subset of the space of potential designs.
Innovative design is less constrained but is still limited to potential designs, with some variables and attributes having extraordinary values.
According to this classification [92], creative design shifts or extends the state space of potential designs.
2.10.2 Design language
In terms of design language, and information models are a design language, routine design combines the available elements in traditional ways. Designing produces an ever more detailed definition of objects, their properties and the relations among them. It assumes modelling software is used as intended, out of the box.
Innovative design extends the predefined object families to new types of objects or enriches existing types of objects with novel properties. It combines types of objects in novel and original
ways. Innovative design requires some programming skills to extend the library of objects out of which designs are created. To do so, expected and predefined features that allow extension and customization of the modelling software are used.
Creative design finds the language offered by the modelling tool, including possible
extensions, insufficient to represent the design ideas. It calls for interpreting existing object types in unintended ways or drops the modelling software altogether. It looks for designs outside of the set that can be represented with a given modelling tool and outside the language of predefined objects.
The latter limitation only exists in software that addresses structured building information and where the design language—the objects that can be used to describe a building—is predefined.
This is not the case if designs are communicated with drawings and lines. Lines are, by definition, interpreted as something else. A line can stand for anything. A wall in a building model can only represent a wall. This can be a weakness.
2.10.3 Design product
The explanations of routine, innovative and creative designs can also be recognized as such when end users experience the product and find it usefully original. A finished building can be found routine, innovative or creative by the general public, regardless of whether the design process or design language fits into the same definition or whether the design language in which it was described before it was built fits the definition of routine, innovative or creative.
The argument that BIM limits creativity is based on the fact that it makes routine designs much easier. Empirical studies could prove (or refute) that designers with BIM tools stick to the elements available, resulting in routine or, at best, innovative designs. In line-based design, there is little technical difference between routine and creative design. In both cases, lines stand for things which they are not. An
argument could be made that stepping out of
16 the usual requires less effort with lines than it does with objects.
However, the above only makes sense if we are interested in intrinsic creativity, i.e., how creative the designer could be and how much freedom or restrictions were imposed by the technology used. The true motive for creativity is not in exercising freedom (which may be more limited in a BIM environment) but in being able to produce better designs, to explore design ideas faster, and to obtain more precise
simulation feedback. Looking at creativity in this way, average designs would be much more creative when using BIM.
A theoretical warning about BIM and creativity, however, remains. Creative designing is
something that philosophers might call “true thinking” [93]. In philosophy, true thinking is defined as “that which is at the edge of what is conceptualized”, i.e., at the edge of what we have words for. Additionally, “true thinking is the creation of concepts” [94]. Designing with BIM uses instances of extremely well-defined concepts. The more BIM is developed, the farther the designer is from that edge. This is a danger to be aware of, as well as an opportunity.
It is the developers of the concepts of BIM that are pushing “the edge of what is conceptualized”
further. Therefore, engineering and architectural creativity is increasing because of the creative work of the developers of BIM software.
3 Discussion and Conclusion
Building information management technology that includes building information modelling tools in the building information modelling process is the most significant technology changing how we design, build, use and manage the built environment. It is a dominant
technological trend in the software industry, and although the theoretical groundwork was laid in the previous century, it is a popular topic in academic research.
In this paper, we have looked at BIM through the lens of structure, function and behavior (Section 2.1). This is the same lens through which BIM is
looking at the built environment. We have defined the structure, function and behavior of BIM, as well as the current state and future development. From a structural and functional perspective, BIM is a database problem. It makes sense to understand it as building information management and the software as building information management systems (BIMS).
Database management systems (DBMS) have been the key tool for the automatization and informatization of many businesses and industries, and BIMS will play the same role in the construction industry (Section 2.2). In fact, the key development problems in BIM structure and function are its (missing) features that databases usually have. It will be harder to address the behavioral issues of how BIMS is affecting its environment and how the environment is affecting BIMS. These social, economic and legal issues will gain prominence as the technical features catch up with other industries.
BIM will contain all the information needed in the building process but will never include the complete information (Section 2.3). This is not a problem because engineers and architects are used to dealing with incomplete information.
The goal of BIM is not to integrate building processes and to reduce the fragmentation of the industry. The true goal, with practical benefits, is to allow even more specialization and division of labor (Section 2.4).
BIM is only starting to change engineering education. Deeper changes will be caused by the redefinition of engineering language and logic.
The deepest changes will be a consequence of automating many easier engineering tasks, restructuring processes and redefining jobs (Section 2.8). Like what happened with management information systems, building information management systems are reducing the flexibility of building processes and are leading to more detailed process definitions.
This calls for a redefinition of how we
understand the activities of the building process (Section 2.3). We suggest the use of game theory
17 and institutional theory to study the adoption of BIM in the industry (Section 2.9). We have argued that BIM standards are not as important as they were in the early days of BIM but remain useful for some many purposes (Section 2.4).
There are serious issues with semantic BIM and semantic interoperability, but the syntactic and structural interoperability around BIM are sufficient (Section 2.6). There are some valid arguments that the intrinsic creativity of engineers and architects who use BIM is reduced; however, there is no doubt that extrinsic creativity is increased by the use of BIM. Those who are truly creative are not only the designers of buildings but the designers of BIM software because they are defining a new language (Section 2.10).
Each of the topics could benefit from a deeper study. Some new angles for such studies were presented, such as a structure-function-behavior view of BIM, understanding BIM as building information management and drawing parallels
[1] Aristotle,
http://www.goodreads.com/quotes/1137536- first-have-a-definite-clear-practical-ideal-a-goal- an, accessed 1.5.2016.
[2] Shakespeare, Henry IV, part 2 | Act 1, Scene 3.
[3] Priest, Susanna Hornig, ed. Encyclopedia of science and technology communication. Sage Publications, 2010.
[4] Henderson, Kathryn. On line and on paper:
Visual representations, visual culture, and computer graphics in design engineering.
Cambridge, MA: MIT Press, 1999.
[5] Björk, Bo-Christer. "RATAS project-developing an infrastructure for computer-integrated construction." Journal of Computing in Civil Engineering 8, no. 4 (1994): 401-419.
[6] Teicholz, Paul, and Martin Fischer. "Strategy for computer integrated construction technology."
Journal of Construction Engineering and Management 120, no. 1 (1994): 117-131.
[7] Wagter, Harry, ed. Computer integrated construction. Elsevier, 2012.
[8] Christiansson P, Karlsson H (ed.); Conceptual modelling of buildings. Proceedings of CIB W74+W78 seminar, October, 1988. Lund
with management information systems,
abandoning the process view of activities, relying on game and institutional theory when studying BIM adoption, the distinction between intrinsic and extrinsic creativity and the creativity of tool makers and tool users.
Perhaps the most important change of perspective of BIM is that it is not a tool of automation or integration but a tool of further specialization. Specialization is a key to the division of labor, which results in greater knowledge, higher quality and more creativity.
Acknowledgements
The research work presented in this paper was funded in part by the Slovenian Research Agency. Their support is gratefully acknowledged.
References
18 university and the Swedish building centre. CIB proceedings 126.
[9] Gu, P., and Kam Chan. "Product modelling using STEP." Computer-Aided Design 27, no. 3 (1995):
163-179.
[10] Krause, F-L., Fumihiko Kimura, T. Kjellberg, SC-Y.
Lu, L. Alting, H. A. Elmaraghy, W. Eversheim et al. "Product modelling." CIRP Annals-
Manufacturing Technology 42, no. 2 (1993):
695-706.
[11] Finger, Susan, and John R. Dixon. "A review of research in mechanical engineering design. Part II: Representations, analysis, and design for the life cycle." Research in Engineering Design 1, no.
2 (1989): 121-137.
[12] Gu, P., and Kam Chan. "Product modelling using STEP." Computer-Aided Design 27, no. 3 (1995):
163-179.
[13] Pratt, Michael J. "Introduction to ISO 10303—
the STEP standard for product data exchange."
Journal of Computing and Information Science in Engineering 1.1 (2001): 102-103.
[14] Reed, Kent A. "Product modeling of buildings for data exchange standards: from IGES to
PDES/STEP and beyond." Conceptual modelling of buildings, CIB W 74 (1988).
[15] Stone, W., K. Reed, P. Chang, L. Pfeffer, and A.
Jacoff. "NIST research toward construction site integration and automation." Journal of aerospace engineering 12, no. 2 (1999): 50-57.
[16] Laiserin, J. (2002): ‘Comparing Pommes and Naranjas‘, The LaiserinLetter, December 16, Issue 15,
http://www.laiserin.com/features/issue15/featu re01.php, accessed 1.5.2016.
[17] Eastman, Charles, David Fisher, Gilles Lafue, Joseph Lividini, Douglas Stoker, and Christos Yessios. "An Outline of the Building Description System. Institute of Physical Planning." (1974).
[18] Eastman, Chuck, Charles M. Eastman, Paul Teicholz, and Rafael Sacks. BIM handbook: A guide to building information modeling for owners, managers, designers, engineers and contractors. John Wiley & Sons, 2011.
[19] Froese, Thomas. "Construction process technologies: a meta-analysis of Canadian research." Canadian Journal of Civil Engineering 36, no. 3 (2009): 480-491.
[20] Howard, Rob, and Bo-Christer Bjork. "Building information models-experts' views on BIM/IFC
developments." In Proceedings of the 24th CIB- W78 Conference, pp. 47-54. 2007.
[21] Kiviniemi, Arto. "Ten years of IFC-development–
Why are we not yet there." In Keynote lecture at the 2006 Joint International Conference on Computing and Decision Making in Civil and Building Engineering, Montreal, Canada. 2006.
[22] Turk, Žiga. "Construction informatics: Definition and ontology." Advanced Engineering
Informatics 20, no. 2 (2006): 187-199.
[23] Panushev, Ivan. Building information
management system. U.S. Patent Application No 13/022,555, 2011.
[24] Society for Marketing Professional Services,
»Evolution of Technology: Innovations in Building Information Marketing«,
https://onlinelearning.smps.org/products/evolu tion-of-technology-innovations-in-building- information-marketing, accessed 1.5.2016.
[25] National Building Specification, BIM Levels Explained,
https://www.thenbs.com/knowledge/bim- levels-explained, accessed 1.5.2016.
[26] Khosrowshahi, Farzad, and Yusuf Arayici.
"Roadmap for implementation of BIM in the UK construction industry." Engineering,
Construction and Architectural Management 19, no. 6 (2012): 610-635.
[27] NBIMS (2010) National Building Information Modeling Standard,
http://www.wbdg.org/pdfs/NBIMSv1_p1.pdf, accessed 1.5.2016.
[28] Miettinen, Reijo, and Sami Paavola. "Beyond the BIM utopia: Approaches to the development and implementation of building information modeling." Automation in construction 43 (2014): 84-91.
[29] de Kleer, J. and Brown, J. S. (1983),
“Assumptions and Ambiguities in Mechanistic Mental Models,” Mental Models, D. Gentner and A. L. Stevens (Eds.), Lawrence Erlbaum Associates, New Jersey, pp. 155-190.
[30] Gero, J.S. (1990). Design Prototypes: A Knowledge representation Schema for Design, AI magazine, Vol 11, No.4.
[31] Watt, Adrienne. Database Design. 2014, https://opentextbc.ca/dbdesign/ accessed 1.5.2016.
[32] B.-C. Bjork, H. Penttila, A scenario for the development and implementation of a building
