7 Information management

This chapter introduces the general goals of information management and connects them to building representations, semantic types and AECO processes in order to distill the main goals of building information management.

The need for information management

With the information explosion we have been experiencing, it is hardly surprising that IM seems to have become a self-evident technical necessity. Handling the astounding amounts of information produced and disseminated every day requires more robust and efficient approaches than ever. Nevertheless, IM is considered mostly as a means to an end, usually performance in a project or enterprise: with effective IM, one can improve the chances of higher performance. Consequently, IM usually forms a key component of overall management.

This is widely acknowledged in building design management (DM). Even before the digital era, the evident dependence of AECO on information that came from various sources and concerned different but interconnected aspects of a building had led to general agreement that this information and the way it is handled can be critical for communication and decision making. DM often focuses on information completeness, relevance, clarity, accuracy, quality, value, timeliness etc., as prerequisites to enabling greater productivity, improving risk management, reducing errors and generally raising efficiency and reliability. The dependence on information is such that some even go so far as to suggest that DM is really fundamentally about IM: managing information flows so that stakeholders receive the right information at the right time.[1]

In practical terms, however, there is little clarity concerning what should be managed and how. DM sources often merely affirm that information is important and should be treated with care. What makes information usable, valuable, relevant etc. is assumed to be known tacitly. Information is vaguely defined as data in usable form but is also equated to the thousands of drawings and other documents produced during the lifecycle of a building — the carriers of information. If the right document is present, then it is assumed that stakeholders also possess the right information and are directly capable of judging the veracity, completeness, coherence etc. of what they receive. However, equating information with documents not only prolongs outdated analogue practices, it also places a heavy burden on users.

It is arguably typical of AECO and DM that, in the face of operational and especially technical complexity, they invest heavily in human resources. This goes beyond the interpretation of documents in order to extract information; it also extends to the invention of new roles that assume a mix of new and old tasks and responsibilities. So, in addition to project and process managers, one encounters information managers as well as BIM managers and CAD managers, BIM coordinators and CAD coordinators, working together in complex, overlapping hierarchies. These new roles are usually justified by the need for support with new technologies, which may be yet unfamiliar to the usual participants in an AECO project. This, however, increases the distance between new technologies and their real users, limiting learning opportunities and prolonging the treatment of technologies as new and unfamiliar (something that contrasts sharply with what we do in our private encounters with new technologies, as discussed in the section on digitization). Even worse, all these roles increase complexity and reduce transparency by adding more intermediaries in the already multi-layered structure of AECO.

New roles are inevitable with technological innovation. Sometimes they are temporary and sometimes permanent. In the early days of motorcars, for example, chauffeurs were more widely employed to drive them than today. On the other hand, webmasters have become necessary by the invention and popularity of the World Wide Web and remain so for the foreseeable future, despite the growing web literacy among general users. What matters is that any such new roles should be part of a sound and thorough plan of approach rather than an easy alternative to a good approach. A good plan should determine what is needed and why, allowing for increasing familiarity and even proficiency of many users with various technologies, to a degree that, after some point, they might require little day-to-day support. In our case, one may safely expect that AECO professionals will eventually become quite capable not only of using BIM directly but also of coordinating their BIM activities, with little need for technical intermediaries. To achieve this, AECO needs practical familiarization with the new technologies but above all clear comprehension of what these technologies do with information. Based on that, one can develop a sound IM approach that takes into consideration both domain needs and the capacities of digital technologies in order to determine changes in the tasks, responsibilities and procedures of existing AECO roles, and develop profiles for any additional roles.

Information sources

Inclusiveness

IM is both an activity and a well-defined discipline of professionals who support this activity. The discipline of IM has a broad scope and, as a result, is quite inclusive.[2] It pays no attention to issues of representation and accepts as information sources all kinds of documents, applications, services and schemes. This is due to three reasons. Firstly, IM covers many application areas and must therefore be adaptable to practices encountered in any of them. Secondly, in many areas there is a mix of analogue and digital information, as well as various channels. For example, financial client transactions with a shop can involve cash and debit or credit cards, either physically or via the web. IM provides tools for bringing such disparate material together into more coherent forms, ensuring that no outdated or inappropriate information is used and preventing that information is missing, inaccessible or deleted by error. These tools include correlation with contexts (e.g. time series displays relative to other data), classification and condensation (aggregation, totalling, filtering and summarization). Thirdly, IM has a tenuous relation to computerization, often relying on it but also appearing weary of putting too much emphasis on digital technologies as a general solution.

The inclusiveness of IM with respect to information sources means that it may end up not only tolerating the redundancy of analogue and digital versions of the same information but also supporting outdated practices and conventions, even prolonging their life through superficial digitization, on the assumption that the application area wants it. This reduces IM to mere document management, i.e. making sure that the necessary documents are retained and kept available. Such inclusiveness is arguably an easy way out of most domain problems. At present, there may be enough computer power and capacity to store and retrieve any document produced in a project or enterprise — in our case, throughout the whole lifecycle of a building. However, the information explosion of the digital era and big data approaches suggest the opposite: we already need more intelligent solutions than brute force. Can we upscale the haphazard, inclusive recording of the history of a building to all buildings in the world? At this moment, we may have the illusion that we still have control over the huge amounts of information in production and circulation but this is because AECO currently approaches information with respect to the limited demands of normative practices. Beyond these demands, there is already too much information that is ignored, neglected and even discarded. Moreover, new developments like the IoT could change the overall picture soon, as smart things start communicating with each other with great intensity. For AECO this can be quite critical because buildings are among the prime candidates for accommodating a wide range of sensors and actuators, e.g. for controlling energy consumption, ensuring security or regulating air quality to prevent the spread of epidemics.

Structured, semi-structured and unstructured information

BIM is important for IM because it marks a transition not only to symbolic representation but also to holistic, structured information solutions for AECO. With regard to structure, there are three main data categories:

  • Unstructured data are the subject of big data approaches: sensor measurements, social media messages and other data without a uniform, standardized format. Finding relevant information in unstructured data is quite demanding because queries have to take into account a broad range of locations where meaningful data may reside and a wide variety of storage forms (including natural language and images).
  • Semi-structured data are a favourite of IM: information sources with a loosely defined structure and flexible use. Analogue drawings are a typical example: one knows what is expected in e.g. a section but there are several alternative notations and few if any prohibitions concerning what may be depicted and how. IM thrives on semi-structured sources, adding metadata, extracting and condensing, so as to summarize relevant information into a structured overview.
  • Structured data are found in sources where one knows precisely what is expected and where. Databases are prime examples of structured information sources. In a relational database, one knows that each table describes a particular class of entities, that each record in a table describes a single entity and that each field describes a particular property of these entities in the same, predefined way. Finding the right data in a structured source is therefore straightforward and less challenging for IM.

In contrast to analogue drawings, BIM is clearly structured, along the lines of a database. Each symbol belongs to a particular type and has specific properties. This structure is one of the driving forces behind BIM, in particular with respect to its capacity to integrate and process building information coherently. Given the effort put into developing structured models in BIM, it makes little sense to abandon the advantages they promise. This makes BIM the main environment for IM in AECO and calls for approaches that should:

  • Avoid having other primary information sources next to BIM. All building information should be integrated in BIM and any related data linked to it. Currently, there is general agreement that the price of a component, e.g. a washbasin, should be a property of the corresponding symbol. However, the same should apply to all data relevant to AECO, e.g. packaging information for this component. The dimensions of the box in which the washbasin is brought to the building site, the packaging materials it contains etc. are useful for logistic purposes, as well as for waste management. Trying to retrieve this information from the manufacturer’s catalogue is significantly less efficient than integrating the relevant data among the symbol properties. The same applies to a photograph of some part of the building during construction or use. This too should be connected to BIM as a link between the digital file of the photograph and relevant symbols in the model (Figure 1) or even mapped as a decal on the symbols (Figure 2).
  • Desist from promoting BIM output to the level of a primary source. Any view of a model, from a floor plan to a cost calculation, can be exported as a separate document (PDF, spreadsheet etc.). Such an export may have its practical uses but one should not treat it as a source separate from the model. Any query about the building should start from the model, bypassing exports and similar output. Using IM to ensure consistency between exports and the model is meaningless. This applies even to legally significant documents like contracts because these too can be expressed as views of the model (i.e. textual frames around data exported from the model).

 

Figure 1. Photograph of current state linked as image to relevant components in Revit

 

Figure 2. Photograph of current state mapped as decal in Revit

 

From the above, a wider information environment emerges around the model, populated largely by data linked to the model, preferably to specific symbols. IM can assist with the organization of this environment but it should not be allowed to cut corners, e.g. answer queries on the basis of satellite files. IM reliability depends on transparent links between queries, external files and the model, specifically the primary data in symbols and their history.

It is perhaps ironic that while the world is focusing on big, unstructured data, AECO should insist on structured data. One explanation is latency: AECO has been late with the development of structured information solutions because it continued to use analogue, semi-structured practices in digital facsimiles. As a consequence, AECO has yet to reap the benefits of structured data approaches, let alone find their limits.

The emphasis on the structured nature of BIM also flies in the face of IM and its inclusiveness. In this respect, one should keep in mind what was discussed in a previous section: IM is a means, not an end, and its adaptability has historical causes. It is not compulsory to retain redundant information sources next to BIM, simply because IM can handle redundancy and complexity. If the structured content of BIM suffices, then IM for AECO simply becomes easier and parsimonious.

Information management goals

Information flow

What we should learn from IM is that the treatment of information should have clear goals. The first of the two main goals of IM is to regulate information flows. This is usually achieved by specifying precise processing steps and stages, which ensure that information is produced and disseminated on time and to the right people, until it is finally archived (or disposed of). In terms of the semantic information theory underlying our approach, this involves identifying and tracking information instances throughout a process, covering both the production and modification of data. IM puts emphasis on the sources and stores of information: the containers from which information is drawn, in which it rests or is archived. BIM combines all these into a single information environment, shifting attention to the symbols, their properties and relations, where all data are found.

Managing information flow involves:

  • What: the information required for or returned by each specific task in a process
  • Who: the actors or stakeholders who produce or receive the information in a task
  • How: the processing of information instances
  • When: the timing of information instances

What is about the paradigmatic dimension: symbols in BIM and external sources linked to them. For both internal and external information, it is critical to distinguish between authorship and custodianship: the actors who produce some information are not necessarily the same stakeholders who safeguard this information in a project, let alone in the lifecycle of a building. A typical example is the brief: this is usually compiled in the initiative stage by a specialist on the basis of client and user input, as well as professional knowledge. In the development stage, custodianship often passes on to a project manager who utilizes the information in the brief to guide and evaluate the design, possibly also adapting the brief on the basis of insights from the design. Then in the use stage, it becomes background to facility and property management, before it develops into a direct or indirect source for a new brief, e.g. for the refurbishment of the building. Making custodianship specific and unambiguous in all stages is of paramount importance in an integrated environment like BIM, where overlaps and grey areas are easy to develop.

How information flows are regulated relates to the syntagmatic dimension of a model: the sequence of actions through which symbols, their properties and relations are processed. The information instances produced by these actions generally correspond to the sequence of tasks in the process but are also subject to extrinsic constraints, including from the software (the implementation environment): the presence of bounding walls is necessary for defining a space in most BIM editors, although in many design processes one starts with the spatial design rather than with construction. IM needs to take such conflicts into account and differentiate between the two sequences.

A useful device for translating tasks into information actions is the tripartite scheme Input-Processing-Output (IPO) that underlies any form of information processing. For any task, some actors deliver information as input. This input is then processed by other (or even the same) actors, who return as output some other information. Then, this output usually becomes input for the next task. IM has to ensure that the right input is delivered to the right actors and that the right output is collected. By considering each decision task with respect to I‑P‑O, one can identify missing information in the input and arrange for its delivery.

The syntagmatic dimension obviously also relates to when: the moments when information instances become available. These moments usually form a coherent time schedule. The time schedule captures the sequence of actions and transactions, linking each to specific information instances. Here again one should differentiate between the sequence of tasks, which tends to be adequately covered by a project schedule, and the sequence of information actions, which may require additional refinement. This difference is the subject of the next part in this book.

Information flow in BIM

We are used to viewing the early part of a design process as something almost magical: someone puts a few lines on a scrap of paper and suddenly we have a basis for imagining what the building will look like. The same applies to BIM: one starts entering symbols in a model and suddenly the design is there for all to see and process. Building information flows seem to emerge out of nothing but this is far from true. The designers who make the first sketches or decide on the first elements in a model operate on the basis of general knowledge of their disciplines, more precise knowledge of the kind of building they are designing and specific project information, including location characteristics and briefs. In other words, building representations are the product of cognitive processes that combine both tacit and overt information.

It is also widely assumed that the amount of information in a design process grows from very little in early design to substantial amounts by the end, when a building is fully specified. This actually refers to the specificity of conventional building representations, e.g. the drawing scales used in different design stages. In fact, even before the first sketch is made, there usually is considerable information available on the building. Some of it predates the project, e.g. planning regulations and building codes that determine much of the form of a building and key features of its elements, such as the pitch of the roof and the dimensions of stairs. Other information belongs to the project, e.g. the brief that states accommodation requirements for the activities to be housed in the building, the budget that constrains cost and location-related principles like the continuation of vistas or circulation networks through the building site. Early building representations may conform to such specifications but most related information remains tacit, either in other documents or in the mind of the designers. For example, the site layout on which one starts drawing or modelling rarely includes planning regulations, even though the designers are normally aware of these regulations and their impact on the design.

In managing both AECO processes and information, one should ensure that tacit information becomes explicit and is connected to tasks. In BIM, this means augmenting the basic model setup (site plan, floor levels, grids etc.) with constraints from planning regulations (e.g. in the form of the permissible building envelope), use information from the brief and constraints on the kind of building elements that are admissible in the model (e.g. with respect to the fire rating of the building). Integration of such information amounts to feedforward: measurement and control of the information system before disturbances occur. Feedforward is generally more efficient and effective than feedback, e.g. checking if all building elements meet the fire safety requirements after they have been entered in the model.

It has also been suggested that early design decisions have a bigger impact on the outcome of a design process than later decisions. Having to decide on the basis of little overt information makes these decisions difficult and precarious. This conventional wisdom concerning early decisions may be misleading. Admittedly, early design decisions tend to concern basic features and aspects, from overall form to load-bearing structure, which determine much of the building and so have a disproportionate influence on cost and performance. However, such decisions are not exclusive to early design: the type of load-bearing structure can change late in the process, e.g. in relation to cost considerations, procurement or the unanticipated need for larger spans. Late changes can be even more expensive because they also necessitate careful control of all interfacing between load-bearing and other elements in the design. Moreover, small, local decisions can also be critical, whether in an early or late stage: if some doors in a building are too narrow, wheelchair circulation may become cumbersome or even impossible, leading to costly restrictions or adaptations. From an IM perspective, what matters is that all relevant information is made explicit in BIM, so as to know which data serve as input for a task and how to register the output of the task. Explicitness of information allows us to map decision making in a process and understand the scope and significance of any decision, regardless of process stage.

Information quality

The second main goal of IM is to safeguard or improve information quality.[3] Quality matters to IM in two respects. Firstly, for information utility: information produced and disseminated in a process should meet the requirements of its users. Secondly, concerning information value: information with a higher quality needs to be preserved and propagated with higher priority. IM measures quality pragmatically, in terms of relevance, i.e. fitness for purpose: how well the information supports the tasks of its users. In addition to pragmatic information quality, IM is also keen on inherent information quality: how well the information reflects the real-world entities it represents. It should be noted that IM is not passive with regard to information quality. It can also improve it, both at meta-levels (e.g. by systematically applying tags) and with respect to content (e.g. through condensation).

In both senses, information quality is determined within each application domain. IM offers a tactical, operational and technical framework but does not provide answers to domain questions. These answers have to be supplied by the application environment in order for IM to know which information to preserve, disseminate or prioritize. In our framework, information quality concerns the paradigmatic dimension: the symbols of a representation and their relations. As this dimension tends to be quite structured in symbolic representations, one can go beyond the pragmatic level of IM and utilize the underlying semantic level to understand better how information quality is determined.

The first advantage of utilizing the semantic level lies in the definition of acceptable data as being well-formed and meaningful. This determines the fundamental quality of data: their acceptability within a representation. A coffee stain cannot be part of a building representation but neither can a line segment be part of a model in BIM: it has to be an explicit symbol of something. That symbol may have the appearance of a line segment (i.e. uses the line segment as implementation mechanism, as is the case for a room separation in Revit) but the meaning of the symbol is not inferred by its appearance — quite the opposite: the appearance is determined by the meaning. Any data that do not fit the specifications of a symbol, a property or a relation cannot be well-formed or meaningful in BIM. Such data are indications of low quality that requires attention. If quality cannot be improved, these data should be treated as noise.

Data that pass the fundamental semantic test must then be evaluated concerning relevance for the particular building or project and its tasks. To judge relevance, one needs additional criteria, e.g. concerning specificity. For example, it is unlikely that a model comprising generic building elements is satisfactory for a task like the acoustic analysis of a classroom because the property values of generic elements tend to be too vague regarding factors that influence acoustic performance.

The semantic level also helps to determine information value beyond utility: prioritizing which information should be preserved and propagated depends on semantic type. As derivative data can be produced from primary when needed, they do not have to be prioritized — in many cases, they do not have to be preserved at all. Operational data and metadata tend to change little and infrequently in BIM, so these too have a lower priority than primary data. Finally, anti-data have a high priority, both because they necessitate interpretation and action, and because such action often aims at producing missing primary data.

Parsimonious IM concerning information quality in a symbolic representation like BIM can be summarized as follows:

  1. Preservation and completion of primary data
  2. Establishing transparent and efficient procedures for producing derivative data when needed
  3. Identification and interpretation of anti-data, including specification of consequent actions
  4. Preservation of stable operational and metadata

The priority of primary data seems to conflict with IM and its improvement of information quality through condensation, i.e. operations that return pragmatically superior derivative data and metadata. Such operations belong to the second point above: if the primary data serve as input for certain procedures, then these procedures have to be established as a dynamic view or similar output in BIM. If users need to know the floor areas of spaces, one should not just give them the space dimensions and let them work out the calculations themselves but supply instead transparent calculations, organized in a legible and meaningful way. This does not mean that the results of these calculations should be preserved next to the space dimensions from which they derive.

Moving from the semantic to the pragmatic level, veracity is a key criterion of quality: fitness for purpose obviously requires that the information is true. In addition to user feedback, veracity can be established on the basis of comparison to additional, reference data, e.g. laser scans that confirm that a model represents faithfully, accurately and precisely the geometry of an existing building.

Before relevance or veracity, however, one should evaluate the structural characteristics of primary information: a model that is not complete, coherent and consistent is a poor basis for any use. Completeness in a building representation means that all parts and aspects are present, i.e. that there are no missing symbols for building elements or spaces in a model. BIM software uses deficiency detection to identify missing symbols. Missing aspects refer to symbol properties or relations: the definition of symbols should include all that is necessary to describe their structure, composition, behaviour and performance.

Completeness is about the presence of all puzzle pieces; coherence is about how well these pieces fit together to produce a seamless overall picture. In a building representation this primarily concerns the interfacing of elements, including possible conflicts in space or time. Clash detection in BIM aims at identifying such conflicts, particularly in space. Relations between symbols are of obvious significance for coherence, so these should be made explicit and manageable.

Finally, consistency is about all parts and aspects being represented in the same or compatible ways. In a symbolic representation, this refers to the properties and relations of symbols. If these are described in the same units and are present in all relevant symbol types, then consistency is also guaranteed in information use. Colour, for example, should be a property of the outer layer of all building elements. In all cases, the colour should derive from the materials of this layer. This means that any paint applied to an element should be explicit as material with properties that include colour. Moreover, any colour data attached to this material layer should follow a standard like the RAL or Pantone colour matching systems. Allowing users to enter any textual description of colour does not promote consistency.

Key Takeaways

  • IM is more than a technical necessity: it is also a means of improving performance in a project or enterprise and therefore a key component of overall management
  • IM is inclusive and accepts all kinds of information, from structured, semi-structured and unstructured sources
  • As a structured information system, BIM simplifies IM
  • IM has two main goals: regulate information flow and safeguard or improve information quality
  • Custodianship of information is critical for information control
  • Information flow relates to the syntagmatic dimension of a representation and draws from the sequence of tasks in a process, as well as from extrinsic constraints
  • In managing information flow one needs to make explicit what, who, how and when
  • The I‑P‑O scheme helps translate tasks into information actions
  • Even before a design takes shape, there are substantial amounts of information that should be made explicit in a model as feedforward
  • Information quality concerns the paradigmatic dimension and can therefore build on the semantic typology of data
  • In addition to semantic and pragmatic criteria, information quality also depends on completeness, coherence and consistency

Exercises

  1. Use the I‑P‑O scheme to explain how one decides on the width of an internal door in a design. Cluster the input by origin (general, specific, project) and describe the relations between input items.
  2. Use the I‑P‑O scheme to explain what, who, how and when in deciding the layout of an office landscape, particularly:
    1. Which workstation types are to be included, including dimensions and other requirements.
    2. How instances of these types are to be arranged to achieve maximum capacity.
  3. In a BIM editor of your choice make the permissible building envelope for a building in a location of your choice. Describe the process in terms of input, information instances produced and resulting constraints for various kinds of symbols in the model.
  4. Evaluate the completeness, coherence and consistency of the permissible building envelope model you have made.
  5. Analyse how one should constrain types of building elements in relation to performance expectations from the use type of building: compare a hotel bedroom to a hospital ward on the basis of a building code of your choice. Explain which symbol properties are involved and how.

  1. The views on DM derive primarily from: Richards, M., 2010. Building Information Management – a standard framework and guide to BS 1192. London: BSI; Eynon, J., 2013. The design manager’s handbook. Southern Gate, Chichester, West Sussex, UK: CIOB, John Wiley & Sons; Emmitt, S., 2014. Design management for architects (2nd ed.). Hoboken NJ: Wiley
  2. The presentation of IM is based on: Bytheway, A., 2014. Investing in information. New York: Springer; Detlor, B., 2010. Information management. International Journal of Information Management, 30(2), 103-108, doi:10.1016/j.ijinfomgt.2009.12.001; Flett, A., 2011. Information management possible?: Why is information management so difficult? Business Information Review, 28(2), 92-100, doi:10.1177/0266382111411066; Rosenfeld, L., Morville, P., & Arango, J., 2015. Information architecture: for the web and beyond (4th ed.). Sebastopol CA: O’Reilly Media.
  3. IM definitions of information quality derive from: Wang, R.Y., & Strong, D.M., 1996. Beyond accuracy: what data quality means to data consumers. Journal of Management Information Systems, 12(4), 5-33. doi:10.1080/07421222.1996.11518099; English, L.P., 1999. Improving data warehouse and business information quality: methods for reducing costs and increasing profits. New York: Wiley.

License

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Building Information - Representation & Management Copyright © by alexanderkoutamanis is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book