The need for classification is not new; we use it almost constantly, daily. But it has one troublesome hitch: we create or shape the classification system in our own way. And the same applies to the construction industry, today almost a large amount of construction information is created digitally. But different in structure and character for each construction profession, each organization, each software tool, or type of actor. It isn’t easy to share this digital information across the lifecycle and across all construction professions, nor can it be used for machine processing. We have no guarantee that different machines will interpret the same information in the same way. However, that’s precisely what we require – the information must be structured and repeatable so that it can be machine-recycled and shared across different tools and platforms.
To communicate in real life, people need to speak the same language. During the construction process, this is very much the same. That’s why we need to agree on a common digital language (CDL). An important component of this common digital language is the classification system. The CDL consists of three basic components and ensures that we recognize, distinguish, and understand the same things in the same way. To make it a national or even an international common digital language, it must meet several basic conditions. First, it must be based on a data dictionary based on the international technical standards EN 23386 and EN 23387. This will ensure that it consists of precise and unambiguous definitions of all building objects and their data templates and properties. These properties must be compiled or determined based on the LOIN (Level of Information Need) principle. This is described in the EN 17412 series of standards. These standards classify the individual properties, and the data templates themselves according to the different actors who require and supply certain information, the reason for which the information is needed and also for a certain milestone when it is needed. A data dictionary must not be just a layer of definitions without specifying – who, why and when the information is required. Thanks to a classification system, thousands of different data templates can be organized in a data dictionary. Thus, the dictionary can clearly cover the actual needs of all actors throughout the life cycle of all different types of buildings.
At the same time, the CLD must be based on open formats that facilitate interoperability. In the case of the construction industry, this is the IFC format, based on ISO 16 379. However, it is far from being the only one; this also applies to the format for the transmission of EPDs, information requirements, budgets, and many others.
The classification system allows us to link machine understanding of building objects with human perception. It is thus an important interface between machine-readable data and information that humans can work with. Moreover, it gives us the ability to group, filter and selectively work with only the subset of the information we require. It is only by composing these three basic components (data dictionary, open formats, and unified classification system) according to international standards that a truly common digital language (CDL) can be created. Thanks to this, information can be maintained, shared and transferred in the long term by different software tools, and by different construction professions (actors) throughout the construction life cycle. Moreover, the use of a common international classification system such as CCI breaks down local limits of cooperation in the construction sector.
The Common Digital Language (CDL) – a path to Understanding
Digital Age Classification System
The traditional classification systems used in the construction industry have been around for many years. So, why change them? But digitalization is significantly changing the way we work, and we need a classification system that meets the needs of the digital age. Existing classification systems typically have a rich tree structure – the classification into classes is very detailed, and many classes have even more subclasses. Unfortunately, this leads in numerous instances to the specific classification being dependent on who is doing the classification. At different stages of the construction life cycle, different actors often classify one item differently. This gives rise to various perspectives, which, however, brings with it a rather significant obstacle to the machine processing of information and its sharing across the construction professions and the construction life cycle. Resolving this contradiction is very difficult in traditional classification systems, within a tree structure, we can only ever use one, and one only, feature to determine the class and subclasses. These properties cannot be further combined.
However, the advent of information technology opens up wholly different possibilities. Suddenly, we have, even in the environment of an ordinary company, quite a lot of computing power at our disposal, as well as capacities for storing and transferring data. This allows us to create very intuitive graphical user interfaces. It makes it possible to combine individual properties or views. What this implies is that a modern classification system of the digital age can leave only a relatively subtle basis for classification. In other words, to use such a granularity of classification that all construction professions agree on it throughout the life cycle. In short, so that everyone classifies a given object in the same way. This gives us a fixed, stable value. With its help, we can then link other structured information generated incrementally throughout the construction life cycle.
Thanks to the possibilities of information technology, we can then combine almost any number of defined properties. These are then stored in data templates according to the LOIN (Level Of Information Needed) principle. Each of the actors can thus use the property/view that corresponds to their needs. But in doing so, the CCI classification ensures that one specific code links the different views, which also allows for use in machine information processing.