How to use classification systems in the age of interoperability?
Parallel digitisation efforts have started to emerge regarding the use of properties, especially within the logic of classification systems. This is why, it is important to look at some considerations which will demystify the use of ‘properties’ in classifications and how this relates to the ongoing standardization work at the EU and International level in order to align both approaches.
The very basic purpose of a classification system is to enable anyone to put a label on something in order to know what it is and how to store it. Then to be able to find this same thing by using the label and give it to someone else who would also understand what it is and what to do with it.
ISO 22274 defines a classification system as ‘a systematic collection of classes organized according to a known set of rules, and into which objects may be grouped’. Simply put, a classification system is the means for grouping objects based on common characteristics.
Historically, construction classification systems were developed mainly to serve the need of organizing data for specific documents such as specifications, cost estimates, FM manuals etc. which are foundation for making informed decisions by distinguishing objects in a specific context.
As there is a diversity of reasons for the collection of certain data for construction objects, it is first necessary to define the purpose of the classification system and then to define the set of characteristics which will be used to identify the classes within, most often, a hierarchical structure. For example, OmniClass has defined a generic class ‘23-17 13 00 Windows’ and based on the property ‘material’, they have further specified subclasses using the combination between the property and possible values, i.e. ‘23-17 13 13 Metal Windows’, ‘23-17 13 15 Wood Windows’.
Recently, classification systems are also becoming a part of ‘identification approaches’, which are used to handle construction data even in a more comprehensive way.
As an addition to the grouping of similar subjects, a.k.a. classification of objects, the combination between a classification system, a numerical system, and/or reference designation system allows users to group objects based on similar characteristics and also identify objects’ occurrences.
By joining classification and identification, users are able to handle complexity in project data and documentation, regardless of the view or purpose for an object.
Construction object characteristics or ‘properties’ within classification systems
Because of the emerging need of structuring information not only regarding subjects with ‘type-of’ and ‘part-of’ relations but also object characteristics or ‘properties’, the industry has started the development of a ‘new age’ classification systems. Those classification systems tend to add a ‘property’ level of information within their classification logic. As this approach is quite new, there are many variations between different classification systems.
Classification systems in the age of interoperability
Many classification systems include properties following a general framework described in ISO 12006-2. However, they also combine the general guidance with national frameworks based on cultural, technological, legal and other specificities. Moreover, as the widely-adopted framework in ISO 12006-2 does not include a specific guidance for the use of properties, it allows for their different interpretation in various classification systems. The standard also suggest the use of properties without actually giving a clear idea and reference on how to define a single property and how to relate it to an object class. This means that the data held within classification systems today might be hard to use in a future driven by interoperability.
Let us have an example case scenario of how this can be an issue:
A Procurement specialist has a requirement to comply with a specific value for the property ‘luminance’ regarding a ‘fixed vertical road traffic sign’ which is specified. He/she sends a price offer request which is based on the set requirements to two manufacturers. The first manufacturer sends back an offer which states that the value for the property ‘luminance’ is ‘pass’. The second manufacturer also sends back an offer which states that the value for the property ‘luminance’ is ‘L1’. Are the values wrong? Why are they different for the same property?
Both of the provided offers state a correct value. They both comply with the same harmonised standard for the relevant product, but they are tested against two different test methods which are described in the harmonised standard. This is why there is a need for the common technical language set by standards to be stored within a data dictionary and accessed through a data template. Knowing the exact definition of the two different test method and their relation to the property ‘luminance’ would provide the possibility of actually comparing products and their performances, and will allow for better-informed choices.
There are also difficulties regarding the mapping of data between classifications. There is a variety of classification systems on the market today. They are all structured in different ways depending on the specific domain of knowledge they cover, national legal frameworks, cultural and technological requirements, and many others. This is why mapping two and more classification systems might be a very difficult task. Some mappings will be one to many, many to one, many to none etc.
This is why there, is a need of a common understanding of what is what, i.e. what is a ‘door’, and what properties belong to this ‘door’ and where are they sourced from in order to support the use of classifications in the digital age. Here is where the abovementioned EU and International standardisation work comes in handy.
How classification systems relate to CEN/TC 442 and Data Templates (DTs)?
Regardless if classification systems have properties or not, no matter what their purpose is, they need a stable source of truth that will allow for mapping between all of them. Such a stable ‘core’ is one well defined ‘concept system’. A ‘concept system’ is not an arbitrary collection of terms but a coherent collection of concepts based on the relations established between them and credible sources providing traceable knowledge within a specific domain (ISO 704).
To provide such a ‘concept system’ is the main aim of CEN/TC 442 WG4 Support Data Dictionaries and the newly proposed standards for Data Templates within a data dictionary framework that follows strict processes of property definition and approval by domain experts. Data Templates are well-defined concept systems, free from limitations in terms of area of applicability, providing expert information already existing in standards. Here are the standards that should be taken into consideration.
prEN ISO 23386 (WI=00442007): Building information modelling and other digital processes used in Construction – Methodology to describe, author and maintain properties in interconnected dictionaries
Provides a methodology to define and manage construction object characteristics for digital use. The principle is to connect every characteristic (called ‘property’ in the digital world) to attributes such as the definition derived from a reference standard within a particular local context. The process creates a rigorous system of validation of all digital contents and defines how ‘properties’ and ‘property groups’ shall be established by experts in a data dictionary, as well as how this content shall be mapped to other data dictionaries. The objective is to allow quality information exchange between industry players for multiple uses such as the digital model, also for international trade, and the needs for maintenance.
prEN ISO 23387 (WI=00442010): Data templates for construction objects used during the inception, brief, design, production, operation and demolition of facilities
Part 1: Concepts defining the general structure of data templates
Sets out the general structure that can be used to digitally describe any construction object within the construction works and building services. This structure is called a Data Template and should be based on concepts and the relationships between concepts sourced from a data dictionary. A Data Template is a collection of standard-based properties collected in groups of properties, which can be traced to credible sources such as, for example, harmonised standards under the CPR, and any other European standards defining construction object characteristics.
BIM is a collaborative process and without having a proper and robust concept system, based on credible sources and standardised processes that inform classification systems, communication and data usage is a mess.
In order to overcome such issues there is a need for a firm dialog between standardisation experts involved with CEN TC/442 and organisations creating classification systems. On the one hand, the classifications community can really help the industry by striking a continuing dialogue with high-level standards policy makers to highlight the impacts, benefits and the implications of standardization for their work. On the other hand, standardisation experts can take advantage of taking the specifics of classification work within their scope. Uniting classifications and a standards-based framework could provide a common and scalable approach that could be used by industry and government in a manner that is consistent with their existing business processes and approaches.