Description
The right management of information is crucial in construction projects and has been studied by many researchers. This paper begins by defining and discussing some information management concepts. In particular, it proposes a reflection about the pull-push model for transferring information and links the resulting ideas to some generic issues of information management, namely information overload, information retrieval and information asymmetry.
MANAGEMENT OF INFORMATION FLOWS DURING
CONSTRUCTION PROJECTS
Benoît Otjacques, Pierre Post and Fernand Feltz
Cellule de Recherche, d’Etude et de Développement en Informatique
Centre de Recherche Public –Gabriel Lippmann, Luxembourg
[email protected]
SUMMARY
The right management of information is crucial in construction projects and has been studied by many
researchers. This paper begins by defining and discussing some information management concepts.
In particular, it proposes a reflection about the pull-push model for transferring information and links
the resulting ideas to some generic issues of information management, namely information overload,
information retrieval and information asymmetry. On the basis of these theoretical elaborations, it
presents a prototype tool for construction project management that is based on original technologies.
Instead of setting up a project specific web site, the approach consists of mixing pull and push
transfer modes on the basis of an enhanced messaging system. This prototype may thus be
considered as a proof-of-concept approach that new technological solutions may be explored if the
conceptual analysis of generic problems takes precedence over default choices in terms of
technology.
INTRODUCTION
Nowadays, information management is known to be one of the most important competitive
advantages of the enterprises. In the construction industry, the design and computing tasks have
been the first ones to intensively use the new information technologies (IT). Using IT to support the
project management and the company organization has emerged as new trend a few years ago. It
has been widely promoted since the Internet widespread all over the world. Therefore, the time has
come to identify the new challenges raised by this paradigm shift.
This paper is organized as follows. First, the pull and push modes of information dissemination are
defined and discussed. Second, some generic information management issues are identified and put
in relation with the dissemination modes. Then, the BBeLink2 prototype is described and confronted
with the previously mentioned issues. The last part proposes some brief conclusions.
MODES OF INFORMATION DISSEMINATION
Most of the time, the discussion about the solutions for exchanging information focuses on
technology. But, in fact, technology isn’t the main point. What really imports is functionality from the
user perspective. Therefore, the basic discussion concerns providing the right information to the right
person at the right moment rather than opposing web site and e-mail systems.
In this paper, the issue of information dissemination is defined as the set of all elements related to
providing a given person with some pieces of information. According to this definition, the information
consumer must be considered as the central element of the reflection and it appears essential to
study the instantiation of the information flow. This naturally brings us to discuss the well-known
‘push-pull’ model of information transfer.
First of all, it is worth reminding some definitions proposed in the literature. The push mode is defined
as follows:
• ‘Data items are sent from the server to the clients without requiring a specific request from the
clients.’ (Acharya et al., 1997)
• ‘The technical definition of push is any automatic mechanism for getting information off the
web from the users perspective.’ (Cerami, 1998)
• ‘Push simply means that new information is delivered or retrieved automatically from a remote
computer to the PC. Information does not need to be updated manually on a regular basis.’
(Basiel, 1999)
• ‘Push (or "server-push") is the delivery of information on the Web that is initiated by the
information server rather than by the information user or client, as it usually is.’ (Whatis.com)
• ‘Push technology (Webcasting) is the prearranged updating of news, weather, or other
selected information on a computer user's desktop interface through periodic and generally
unobtrusive transmission over the World Wide Web’ (SearchWebServices.com)
Some of these definitions lack of precision concerning at least one fundamental element: the layer to
which they apply. Some relate to the user and others focus on the computer. However, the need for
accurate definitions has been identified for many years. The 7-layers OSI model, underlying almost all
technologies concerned by information transfer, is probably the most important result in this domain. It
permits to clearly identify which nodes are exchanging information and which services are linked to
the transfer.
Similar definitions have been proposed for the pull mode. Nevertheless, for the purpose of this paper,
it is sufficient to mention one of them. ‘Using request-response, clients explicitly request data items by
sending messages to a server. When a data request is received at a server, the server locates the
information of interest and returns it to the client. Pull-based access has the advantage of allowing
clients to play a more active role in obtaining the data they need, rather than relying solely on the
schedule of a push based server.’ (Acharya et al., 1997)
This paper discusses the information management in construction projects. This isn’t a
telecommunication oriented paper. Therefore, the user level is considered as the reference point for
the proposed definitions.
• The push mode concerns any information flow that occurs without explicit request of the
person that receives it. In this case, the receiver plays a passive role.
• The pull mode concerns any information flow that occurs as a consequence of an explicit
request of a person. In this case, the receiver plays an active role.
This can be illustrated by some examples. A phone call is obviously an implementation of the push
mode, as well as a warning window on the computer screen. Similarly, surfing on a web site and
searching a phone book are considered as pull mode.
User
interface
User
interface
pull pull
push push
Figure 1 Push-pull modes from the user perspective
In the domain of information technologies, it must be noticed that a push mode at layer
may be
implemented on a pull mode at the layer (n-1) and vice versa. For instance, while the e-mail
technology is often considered as a push technology, it lays on a connection to a mail server
instantiated by the client application at a lower level (pull mode). It is thus important to specify which
level is concerned. According to the definitions adopted in this paper, the human-computer interface
must be considered to evaluate whether the functionality implements the pull or the push mode,
independently of the underlying technology (cf. Figure 1). It may be useful to illustrate this discussion
with an example. The problems caused by the size of the data to be exchanged are often presented
as intrinsic features of the push or pull mode but, in fact, they rely on the technology used to
implement the chosen mode instead of the mode itself. One shouldn’t confuse the assessment of
some communications technologies such as SMTP, HTTP or ISDN with some conceptual reflection
about the information dissemination.
This distinction between the pull and push modes is not only a theoretical debate but underpins some
critical issues from a management viewpoint that are discussed in the next section.
THE INFORMATION MANAGEMENT ISSUES
The effective management of information is known to have a critical influence on the performance of a
company. Previous research focusing on the construction industry have shown that this assertion is
also valid in this domain. During a construction project, hundreds of information flows occur. The high
number of involved companies makes the effective management of these exchanges crucial. Indeed,
the steps of the production process are highly interrelated (Koskela and Vrijhoef, 2000), which leads
to a strong dependency among the companies engaged in the project. Suppliers commonly contribute
75-80% of the value of a construction contract, making their effective management and coordination
essential to cost, quality and time objectives (Clark et al., 1999). The problems related to the
transmission of information have been identified as very important factors of inefficiency in a
construction project. For instance, according to O’Connor and Tucker (1999), 22,1% of the problems
occurring on a construction site relate to the communication of design information. It has also been
estimated that ‘up to 30% of the cost of a building project is due to the fractured processes and
communication of the AEC/FM industry’ (I.A.I. - International Alliance for Interoperability, 2002). To
sum up, a lack of information flow management often leads to delays, additional costs and
consequently bad relationships with the customers. It is thus founded to discuss in more detail some
generic issues that face information management in the construction sector.
Information overload
The first problem to discuss is the information overload. It is often cited as a major problem for the
managers in general (Farhoomand and Drury, 2002) and the construction professionals in particular
(Thorpe and Mead, 2001). This issue arises in two main contexts: ‘when the user are given more
information than they can absorb’ and ‘when the information processing demand on an individual’s
time for performing interactions and internal calculations exceeds the supply or capacity of time
available for such processing’ (Farhoomand and Drury, 2002). The former case relates to the amount
of information to be processed and the latter to the time available to information processing.
Information retrieval
A second point worth being studied is the issue of information retrieval. This is linked to the
information structuring and ease of access. When some information pieces are stored in a repository,
the extent to which they are structured and indexed gains a fundamental importance as it underpins
the performance of the search process. Moreover, the structure widely influences the degree to which
some information processing tasks may be automated. Finally, ergonomics also plays a major role in
this domain by influencing the cognitive effort required by the user to get access to the information.
Information asymmetry
Information asymmetry arises when all the persons dealing with a given problem have different levels
of information about the same object. This difference may take several forms:
• distortion, due to a loss of information integrity during the communication,
• incompleteness, due to a loss of information pieces during the communication.
In most of the cases, the asymmetry results from the way that information is managed within the
group of involved persons. This paper doesn’t consider the asymmetry as the consequence of a
deliberate behaviour of some members. The asymmetry may thus be caused by a deficient
synchronization of information within the group. In this case, the members have a version of the
information that corresponds to different moments of time. The most usual solution to tackle this
problem is to set up a reference information repository that is hypothesized to contain the latest
version of the information pieces. Consequently, in this context, the refresh rate of the data taken into
account by the user or, in others terms, the frequency of the connection to the reference repository,
becomes the central element. The asymmetry may also originate from an imperfect transmission path
that modifies the integrity and/or the completeness of the content. In this case, the asymmetry is
mainly influenced by the choice of the information transfer mode. Indeed, it is well known that the
higher the number of intermediaries in the transmission path the higher the probability of information
corruption. In such a situation, the solution consists of providing a mean to certify that the information
hasn’t been corrupted during the transfer.
RELATION BETWEEN ISSUES AND DISSEMINATION MODES
On one hand, it has previously been shown that two basic modes of information dissemination
coexist. On the other hand, three fundamental issues have been identified in the domain of
information management. This section explains the relations that link these observations together.
Information
dissemination modes
Push
Pull
Information management
issues
Overload
Retrieval
Asymmetry
?
Figure 2 Relation between information management and dissemination modes
The push mode is very useful when the sender demands the receiver to react to the incoming
information. The push mode means that some stimuli are sent to the receiver and force him to adopt a
given behaviour (read – ignore – answer). This mode is therefore the most adapted to handle
information flows that require a rapid answer (e.g. phone call to warn the company that an accident
occurred on the construction site). It is also useful to deal with events having some unexpected
components (e.g. invitation for a meeting). The push mode also suits the implementation of reminding
messages, such as those used in most of the calendars. Finally, it may also be used to tell the user
that some elements of the context have changed (e.g. service to alert the users when data has been
added or modified in a web site).
The major drawback of the push mode concerns the information overload issue. If a large amount of
information is sent to the receiver, it may exceed his cognitive capacity (e.g. spam using e-mail).
Therefore, an efficient push implementation must filter the communications in order to keep the
number of information flows at an acceptable level. An important remark must be made in this
context. In many circumstances, some incoming messages pushed to the user are archived for later
use. In this case, according to the definition adopted in this paper, the retrieval of these information
pieces must be considered as pull mode (causing the usual problems associated to this configuration,
see hereafter). This observation illustrates the fundamental importance to clearly identify the layer
considered in the discussion to avoid confusion between technology and functionality. Another
drawback of the push mode relates to asymmetry. It is important to notice however that the potential
importance of the asymmetry issue depends on the way how the push mode is used in practice. If the
information is pushed from point to point within the group, some kinds of asymmetry appear quite
irremediably either due to the transmission delay among the nodes (cf. deficient synchronization) or
due to content modification by one of them (cf. imperfect transmission). If one trustworthy information
provider sends information to several persons via the push mode (i.e. multicasting), the imperfect
transmission problem may be avoided and it only remains the synchronization issue.
Finally, if a viewpoint that is consistent to the proposed definitions is adopted, the push mode isn’t
involved by the retrieval issue because the communication is instantiated by the computer.
The pull mode permits to limit the information overload because the user doesn’t have to handle any
disturbing incoming information flow. The information is picked up from a repository when needed.
The pull mode suppresses the well-known issue of ‘just-in-case-they-need-it’ push publishing (Thorpe
and Mead, 2001) (which must be considered, in fact, as a misuse of the push mode). The greatest
advantage of the pull mode is that it doesn’t offer the user an easy way to abusively trigger hundreds
of information flows. There is however a counter part of this situation: the dependence on the initiative
taken by the user. The pull mode isn’t able to resolve a situation where a project participant doesn’t
instantiate any connection to the repository.
The pull mode presents also an advantage in terms of information symmetry. Because one
information source is hypothesized to be up-to-date, it is easy to know where to collect the latest
version of any information piece. As already mentioned, the efficiency of the pull mode in this domain
primarily depends on the refresh rate of the data. More precisely, in most of the pull mode
implementations, a temporary buffer storing previously retrieved information is used to optimize the
performance. Due to this feature, the data may be accessed while the synchronization mechanism is
deactivated and the risk of consulting old data reaches the same level as in the multicasting solution.
The information retrieval issue must probably be considered as the major drawback of the pull mode.
As the amount of information to be searched increases, the difficulty to rapidly find the required data
becomes more important. In the worst case, the required effort may discourage the person to search
the repository and consequently to store anything into. Therefore, an efficient pull mode
implementation must provide structured information combined with user-friendly retrieval tools.
THE SITUATION IN THE CONSTRUCTION INDUSTRY
In practice, most of the solutions proposed to manage construction projects combine some push and
pull functionalities. This can be illustrated with the example of a project specific web site (PSWS),
which is nowadays probably the most promoted approach. While this solution is often presented as
the perfect example of pull mode, it includes, most of the time, some push-based functionalities (e.g.
e-mail technology used for data update notification or invitation to meetings). The push-oriented
approaches also often include some pull functions. For instance, an electronic message sent to invite
the receiver to visit a website may be qualified in this way. The effective management of a project
requires thus an environment able to deal with push and pull communications.
Instead of improving pull-based project management environment with some push functionalities, we
have chosen to adopt the symmetric way: designing an enhanced push-based tool – the BBeLink2
prototype – including some pull features. This approach is based on the observation that web sites
present widely discussed advantages but also some drawbacks pointed out by quite few authors.
Most of the web site systems face the same problems than those encountered for years by integrated
project databases (IPDB). For instance, information retrieval becomes more difficult as the amount of
data increases, data partitioning and access rights may be very difficult to manage (see Amor & Faraj,
2001, for a comprehensive discussion on misconceptions about IPDB). Moreover, they cause the
emergence of new kinds of risks for the companies, especially for the smallest ones:
• dependence on the (potentially rival) company hosting the web site,
• dependence on the usage by all projects participants,
• confidentiality of the data,
• risk of failure of the web site operator.
Finally, the greatest challenge for the PSWS is probably to bring the construction professionals to
replace some of the traditional point-to-point communications by the consultation, relying on personal
initiative, of a central repository. The effective use of a PSWS requires the project participants to
change their way of performing some of their tasks. Unfortunately, some recent empirical experiences
have concluded that ‘even six years is a short time to change human attitudes to the means of
communication used, in spite of impressive new technologies being available.’ (Howard and Petersen,
2001) In conclusion, ‘the organisational issues surrounding the use as well as the psychology
involved in getting all participants in projects to accept using new technology are now in focus.’ (Björk,
2002)
From this perspective, the BBeLink2 approach is less ambitious. It is based on the strategy ‘which
does not present technology as a new mode of organisation but only as the way to introduce its
necessity’ and this approach ‘is not unfounded and has been successfully used within other industries
[than the construction]’. (Brousseau and Rallet, 1993). This approach also takes into account the
importance of proposing technologies that may be rapidly accepted, which has been recognized as a
fundamental element of acceptation for the small companies (Mundim and Bremer, 2000).
BBeLink2 has been designed to handle both point-to-point communications like classic e-mails and
multicasting dissemination, without using a central web site. The multicasting diffusion of information
may be instantiated only by the project creator. It is used to diffuse global project properties that
present some interest for every participant (e.g. picture of the current state of the building site, project
participant list, building description). This feature is intended to increase the confidence of every
participant in the data validity. A technological description is provided in the next section.
BBELINK2 TOOL
Technology description
The BBeLink2 prototype is based on the creation of a community. Every member of the system uses
a client application that handles the connections to a server. There are only two centralized modules
shared by all the participants: a communication server used to transfer the messages and a central
repository containing the identification of the people/companies allowed to use the system. There isn’t
any central repository for project data.
BBeLink2 includes two major types of functionalities: communication modules and storage modules.
The communication issues are handled by implementing two kinds of messages that may be
transferred among the client applications: user messages and system messages. The former may be
compared to classic e-mails that would be enhanced by some value-added features (e.g. reliable time
synchronization for message dating, inclusion of meta data describing some message properties).
They are dedicated to be composed and read by human beings. The user messages include some
primary and secondary data. The primary data are those that should be read by the receiver in all
cases (i.e. sender, sent date, subject, content text). The secondary data regroups those that may be
read by the user if he wants to process the message in a comprehensive manner (e.g. for sent
messages: reception status for every receiver; for received messages: attached files). The system
messages are exchanged between computers for performing some low-level tasks in a transparent
manner for the user (e.g. acknowledgement mechanism for every user message, synchronization of
global project data via multicasting).
The storage modules are dedicated to store the information related to the projects. This data
describes two kinds of objects: the communication instances and the project properties. The former
concerns the description of all the information flows (IFs) occurring among the project stakeholders,
independently of the communication media used. BBeLink2 includes a repository for storing all IFs. It
distinguishes a structured and an unstructured part in an IF description. The structured part is
implemented as a set of standardized meta-data to assign to any IF. Some are very usual, such as
the identification of the sender, and others are more original, such as the content type (technological,
administrative, commercial…) or the communication media used (phone, fax, meeting, e-mail…). The
meta-data set has been instantiated from a three-step process. First, a literature review has provided
a list of potential values (e.g. list of project stages). Second, a limited list of values has been built,
which appeared as an acceptable compromise between usability and utility. Third, the resulting list
was revised by a partner enterprise active for several years in the AEC industry. The unstructured
part allows to describe the IF with free text and to attach any kind of file (picture, drawing, contract…).
The IFs linked to the BBeLink2 user messages form a specific subset. Because the user messages
are instantiated inside the BBeLink2 environment, the corresponding IFs are automatically created in
the IFs repository. The latter concerns the project itself and is implemented by project properties (e.g.
building description, construction site picture, participants list). BBeLink2 distinguishes global and
local project properties. The former are defined by the project creator. They are diffused to the project
participants and synchronized by multicasting. The latter are also synchronized but they may be
modified by the project participants.
BBeLink2 and the information management issues
The BBeLink2 concept must then be confronted with the three basic issues related to information
management that have been previously mentioned. In order to facilitate the understanding of the
different discussed elements, the Figure 3 provides a global view of the BBeLink2 features, in terms
of information dissemination modes.
The first issue to analyze is the information overload. BBeLink2 includes several features intended to
handle this issue.
• The server handles only messages of the members of the community. Moreover, it may be
configured to block the messages of any member showing an inappropriate behaviour. This
mechanism may thus limit the risk of information overload due to spam.
• The BBeLink2 clients integrate a filtering mechanism. A filter is a logic expression defined
using primary (preferably) or secondary data of the user messages. A filter may also be a
combination of existing filters. The filters are evaluated for each incoming message. If the
expression is true, a specified action (e.g. forward, save in specified directory…) is
automatically triggered. This mechanism may limit the amount of incoming flows to handle.
• The distinction between primary and secondary data in the user messages also limits the
information to handle during the first step of the cognitive processing. Moreover, the user may
choose which (primary) meta data of the messages should be displayed in the graphical
interface.
• It is possible to generate summary reports about the information flows that have occurred
during the project. This may help the professionals to rapidly get an overview of the current
state of a project.
• The user may deactivate the push-based warning messages sent when some global project
properties are updated.
• The user interfaces used to handle user messages and information flows are very similar, in
order to reduce the learning effort for the user.
Server
{System Messages, User Messages}
User A
Application client A
Project properties
Messages A
Information
flows
A
Global
User Messages A
System Messages A
Local
Primary
data
Secondary
data
Push Pull Pull Pull
Push
(update warning)
C
o
m
m
u
n
i
c
a
t
i
o
n
m
o
d
u
l
e
s
S
t
o
r
a
g
e
m
o
d
u
l
e
s
Figure 3 BBeLink2 conceptual model (push-pull viewpoint)
The second issue concerns the retrieval of information. The design of BBeLink2 has tried to structure
the data as far as possible, in order to ease the information access.
• The whole BBeLink2 environment is organized by project. Every message and every
information flow is linked to a specified project.
• Every user message and information flow is characterized by a (similar) set of meta data. This
allows an efficient sorting and retrieval of any instance of communication.
• The filtering mechanism may be used to automatically organize the incoming user messages
in a given way, specific to the company.
The information asymmetry is the last issue to be discussed.
• The multicasting system guarantees that the global project properties are identical for every
project participant. Indeed, only the project creator is granted to set and modify them. The
local replication that is stored on the computer of each participant is encrypted and may not
be accessed in write mode.
• The information synchronization depends on the connection frequency to the Internet. As
soon as a new connection is instantiated, every new modification of the global project data is
automatically updated on the local instance. This update is transparent to the user and can’t
be deactivated.
• BBeLink2 also provides a reliable dating system. Each time value set in relationship to a
message is collected from an external time server. In conjunction with the low level,
permanently activated acknowledgment mechanism, this permits to be really informed on the
reception status of any message, for any receiver. The sender may thus also be aware of a
lack of synchronization.
• By integrating a digital signature implementation, BBeLink2 prevents the corruption of the
transferred data.
CONCLUSIONS
In many discussions, the pull and push mode are presented as rival modes of information
dissemination while they are in fact complementary. Moreover, the technological viewpoint often
dominates the functional perspective. This paper describes a proof-of-concept tool, called BBeLink2,
which mixes push and pull functionalities. It is original in many ways but the absence of a central web
site is certainly one of the most worthy to be mentioned. The next step aims to get a sample of
architects to evaluate the tool, especially in terms of utility, usability and intend of use.
REFERENCES
Acharya B., M. Franklin and S. Zdonik (1997), Balancing Push and Pull for Data Broadcast in
Proceedings of ACM SIGMOD Conference, May 1997, Tuscon, Arizona.
Amor R. and Faraj I. (2001), Misconceptions about integrated project databases, in Electronic Journal
of Information Technology in Construction,http://www.itcon.org/2001/5/, accessed 12 Dec.
2002.
Basiel A. (1999), Push & Pull technologies using a WIDE approach – Web-based Instructional
Designed Environment, 3rd Human Centred Technologies Postgraduate Workshop 1999,
Brighton.
Björk, B.-C. (2002), The impact of Electronic Document Management on Construction Information
Management, in Proceedings of the CIB w78 Conference, Distributing knowledge in building,
June 2002, Aarhus, Denmark.
Brousseau E. and A. Rallet (1993), Développement des systèmes télématiques et évolution des
relations interentreprises dans la construction, Rapport d’une étude financée par le Plan
Construction et le PIRTTEM, Université Paris-Sud, available on demand fromhttp://www.jm.u-
psud.fr
Cerami E. (1998), Delivering push, Mc Graw Hill, NY.
Clark A. M., B. L. Atkin, M. P. Betts and D. A. Smith (1999), Benchmarking the use of IT to support
supplier management in construction, in Electronic Journal of Information Technology in
Construction,http://www.itcon.org/1999/1/, accessed 12 Dec. 2002.
Farhoomand A. F. and D. H. Drury (2002), Managerial information overload in Communications of the
ACM, October 2002, Vol. 45, No. 10, pp. 127-131.
Howard R. and E. Petersen (2001), Monitoring communications in partnering projects in Electronic
Journal of Information Technology in Construction,http://itcon.org/2001/1, accessed 26 Nov.
2002.
I.A.I. (2002),http://cig.bre.co.uk/iai_uk/iai/page2.htm, accessed 11 Dec. 2002.
Koskela L. and R. Vrijhoef (2000), The prevalent theory of construction is a hindrance for innovation,
in Proceedings of International Group for Lean Construction 8th Annual Conference,http://www.susx.ac.uk/spru/imichair/news/template.cfm?content=iglc-8.cfm, accessed 14 Dec.
2002.
Mundim A. and C. F. Bremer (2000), Design of a computer-supported cooperative environment for
small and medium enterprises, 2
nd
IFIP Conference: Managing Business-to-Business
Cooperation, Dec. 2000, Florianopolis, Brazil.
O’Conner J.T. and R. L. Tucker (1986), Improving Industrial Project Constructability, in Journal of
Construction Engineering and Management, Vol. 112, No1, March, pp 69-82.
SearchWebServices.com (2002),http://searchwebservices.techtarget.com/, accessed 14 Dec. 2002.
Thorpe T. and S. Mead (2001), Project-Specific Web Sites: Friend or Foe? in Journal of Construction
Engineering and Management, September/October 2001.
Whatis.com (2002),http://whatis.techtarget.com/, accessed 14 Dec. 2002.
doc_460838595.pdf
The right management of information is crucial in construction projects and has been studied by many researchers. This paper begins by defining and discussing some information management concepts. In particular, it proposes a reflection about the pull-push model for transferring information and links the resulting ideas to some generic issues of information management, namely information overload, information retrieval and information asymmetry.
MANAGEMENT OF INFORMATION FLOWS DURING
CONSTRUCTION PROJECTS
Benoît Otjacques, Pierre Post and Fernand Feltz
Cellule de Recherche, d’Etude et de Développement en Informatique
Centre de Recherche Public –Gabriel Lippmann, Luxembourg
[email protected]
SUMMARY
The right management of information is crucial in construction projects and has been studied by many
researchers. This paper begins by defining and discussing some information management concepts.
In particular, it proposes a reflection about the pull-push model for transferring information and links
the resulting ideas to some generic issues of information management, namely information overload,
information retrieval and information asymmetry. On the basis of these theoretical elaborations, it
presents a prototype tool for construction project management that is based on original technologies.
Instead of setting up a project specific web site, the approach consists of mixing pull and push
transfer modes on the basis of an enhanced messaging system. This prototype may thus be
considered as a proof-of-concept approach that new technological solutions may be explored if the
conceptual analysis of generic problems takes precedence over default choices in terms of
technology.
INTRODUCTION
Nowadays, information management is known to be one of the most important competitive
advantages of the enterprises. In the construction industry, the design and computing tasks have
been the first ones to intensively use the new information technologies (IT). Using IT to support the
project management and the company organization has emerged as new trend a few years ago. It
has been widely promoted since the Internet widespread all over the world. Therefore, the time has
come to identify the new challenges raised by this paradigm shift.
This paper is organized as follows. First, the pull and push modes of information dissemination are
defined and discussed. Second, some generic information management issues are identified and put
in relation with the dissemination modes. Then, the BBeLink2 prototype is described and confronted
with the previously mentioned issues. The last part proposes some brief conclusions.
MODES OF INFORMATION DISSEMINATION
Most of the time, the discussion about the solutions for exchanging information focuses on
technology. But, in fact, technology isn’t the main point. What really imports is functionality from the
user perspective. Therefore, the basic discussion concerns providing the right information to the right
person at the right moment rather than opposing web site and e-mail systems.
In this paper, the issue of information dissemination is defined as the set of all elements related to
providing a given person with some pieces of information. According to this definition, the information
consumer must be considered as the central element of the reflection and it appears essential to
study the instantiation of the information flow. This naturally brings us to discuss the well-known
‘push-pull’ model of information transfer.
First of all, it is worth reminding some definitions proposed in the literature. The push mode is defined
as follows:
• ‘Data items are sent from the server to the clients without requiring a specific request from the
clients.’ (Acharya et al., 1997)
• ‘The technical definition of push is any automatic mechanism for getting information off the
web from the users perspective.’ (Cerami, 1998)
• ‘Push simply means that new information is delivered or retrieved automatically from a remote
computer to the PC. Information does not need to be updated manually on a regular basis.’
(Basiel, 1999)
• ‘Push (or "server-push") is the delivery of information on the Web that is initiated by the
information server rather than by the information user or client, as it usually is.’ (Whatis.com)
• ‘Push technology (Webcasting) is the prearranged updating of news, weather, or other
selected information on a computer user's desktop interface through periodic and generally
unobtrusive transmission over the World Wide Web’ (SearchWebServices.com)
Some of these definitions lack of precision concerning at least one fundamental element: the layer to
which they apply. Some relate to the user and others focus on the computer. However, the need for
accurate definitions has been identified for many years. The 7-layers OSI model, underlying almost all
technologies concerned by information transfer, is probably the most important result in this domain. It
permits to clearly identify which nodes are exchanging information and which services are linked to
the transfer.
Similar definitions have been proposed for the pull mode. Nevertheless, for the purpose of this paper,
it is sufficient to mention one of them. ‘Using request-response, clients explicitly request data items by
sending messages to a server. When a data request is received at a server, the server locates the
information of interest and returns it to the client. Pull-based access has the advantage of allowing
clients to play a more active role in obtaining the data they need, rather than relying solely on the
schedule of a push based server.’ (Acharya et al., 1997)
This paper discusses the information management in construction projects. This isn’t a
telecommunication oriented paper. Therefore, the user level is considered as the reference point for
the proposed definitions.
• The push mode concerns any information flow that occurs without explicit request of the
person that receives it. In this case, the receiver plays a passive role.
• The pull mode concerns any information flow that occurs as a consequence of an explicit
request of a person. In this case, the receiver plays an active role.
This can be illustrated by some examples. A phone call is obviously an implementation of the push
mode, as well as a warning window on the computer screen. Similarly, surfing on a web site and
searching a phone book are considered as pull mode.
User
interface
User
interface
pull pull
push push
Figure 1 Push-pull modes from the user perspective
In the domain of information technologies, it must be noticed that a push mode at layer
may beimplemented on a pull mode at the layer (n-1) and vice versa. For instance, while the e-mail
technology is often considered as a push technology, it lays on a connection to a mail server
instantiated by the client application at a lower level (pull mode). It is thus important to specify which
level is concerned. According to the definitions adopted in this paper, the human-computer interface
must be considered to evaluate whether the functionality implements the pull or the push mode,
independently of the underlying technology (cf. Figure 1). It may be useful to illustrate this discussion
with an example. The problems caused by the size of the data to be exchanged are often presented
as intrinsic features of the push or pull mode but, in fact, they rely on the technology used to
implement the chosen mode instead of the mode itself. One shouldn’t confuse the assessment of
some communications technologies such as SMTP, HTTP or ISDN with some conceptual reflection
about the information dissemination.
This distinction between the pull and push modes is not only a theoretical debate but underpins some
critical issues from a management viewpoint that are discussed in the next section.
THE INFORMATION MANAGEMENT ISSUES
The effective management of information is known to have a critical influence on the performance of a
company. Previous research focusing on the construction industry have shown that this assertion is
also valid in this domain. During a construction project, hundreds of information flows occur. The high
number of involved companies makes the effective management of these exchanges crucial. Indeed,
the steps of the production process are highly interrelated (Koskela and Vrijhoef, 2000), which leads
to a strong dependency among the companies engaged in the project. Suppliers commonly contribute
75-80% of the value of a construction contract, making their effective management and coordination
essential to cost, quality and time objectives (Clark et al., 1999). The problems related to the
transmission of information have been identified as very important factors of inefficiency in a
construction project. For instance, according to O’Connor and Tucker (1999), 22,1% of the problems
occurring on a construction site relate to the communication of design information. It has also been
estimated that ‘up to 30% of the cost of a building project is due to the fractured processes and
communication of the AEC/FM industry’ (I.A.I. - International Alliance for Interoperability, 2002). To
sum up, a lack of information flow management often leads to delays, additional costs and
consequently bad relationships with the customers. It is thus founded to discuss in more detail some
generic issues that face information management in the construction sector.
Information overload
The first problem to discuss is the information overload. It is often cited as a major problem for the
managers in general (Farhoomand and Drury, 2002) and the construction professionals in particular
(Thorpe and Mead, 2001). This issue arises in two main contexts: ‘when the user are given more
information than they can absorb’ and ‘when the information processing demand on an individual’s
time for performing interactions and internal calculations exceeds the supply or capacity of time
available for such processing’ (Farhoomand and Drury, 2002). The former case relates to the amount
of information to be processed and the latter to the time available to information processing.
Information retrieval
A second point worth being studied is the issue of information retrieval. This is linked to the
information structuring and ease of access. When some information pieces are stored in a repository,
the extent to which they are structured and indexed gains a fundamental importance as it underpins
the performance of the search process. Moreover, the structure widely influences the degree to which
some information processing tasks may be automated. Finally, ergonomics also plays a major role in
this domain by influencing the cognitive effort required by the user to get access to the information.
Information asymmetry
Information asymmetry arises when all the persons dealing with a given problem have different levels
of information about the same object. This difference may take several forms:
• distortion, due to a loss of information integrity during the communication,
• incompleteness, due to a loss of information pieces during the communication.
In most of the cases, the asymmetry results from the way that information is managed within the
group of involved persons. This paper doesn’t consider the asymmetry as the consequence of a
deliberate behaviour of some members. The asymmetry may thus be caused by a deficient
synchronization of information within the group. In this case, the members have a version of the
information that corresponds to different moments of time. The most usual solution to tackle this
problem is to set up a reference information repository that is hypothesized to contain the latest
version of the information pieces. Consequently, in this context, the refresh rate of the data taken into
account by the user or, in others terms, the frequency of the connection to the reference repository,
becomes the central element. The asymmetry may also originate from an imperfect transmission path
that modifies the integrity and/or the completeness of the content. In this case, the asymmetry is
mainly influenced by the choice of the information transfer mode. Indeed, it is well known that the
higher the number of intermediaries in the transmission path the higher the probability of information
corruption. In such a situation, the solution consists of providing a mean to certify that the information
hasn’t been corrupted during the transfer.
RELATION BETWEEN ISSUES AND DISSEMINATION MODES
On one hand, it has previously been shown that two basic modes of information dissemination
coexist. On the other hand, three fundamental issues have been identified in the domain of
information management. This section explains the relations that link these observations together.
Information
dissemination modes
Push
Pull
Information management
issues
Overload
Retrieval
Asymmetry
?
Figure 2 Relation between information management and dissemination modes
The push mode is very useful when the sender demands the receiver to react to the incoming
information. The push mode means that some stimuli are sent to the receiver and force him to adopt a
given behaviour (read – ignore – answer). This mode is therefore the most adapted to handle
information flows that require a rapid answer (e.g. phone call to warn the company that an accident
occurred on the construction site). It is also useful to deal with events having some unexpected
components (e.g. invitation for a meeting). The push mode also suits the implementation of reminding
messages, such as those used in most of the calendars. Finally, it may also be used to tell the user
that some elements of the context have changed (e.g. service to alert the users when data has been
added or modified in a web site).
The major drawback of the push mode concerns the information overload issue. If a large amount of
information is sent to the receiver, it may exceed his cognitive capacity (e.g. spam using e-mail).
Therefore, an efficient push implementation must filter the communications in order to keep the
number of information flows at an acceptable level. An important remark must be made in this
context. In many circumstances, some incoming messages pushed to the user are archived for later
use. In this case, according to the definition adopted in this paper, the retrieval of these information
pieces must be considered as pull mode (causing the usual problems associated to this configuration,
see hereafter). This observation illustrates the fundamental importance to clearly identify the layer
considered in the discussion to avoid confusion between technology and functionality. Another
drawback of the push mode relates to asymmetry. It is important to notice however that the potential
importance of the asymmetry issue depends on the way how the push mode is used in practice. If the
information is pushed from point to point within the group, some kinds of asymmetry appear quite
irremediably either due to the transmission delay among the nodes (cf. deficient synchronization) or
due to content modification by one of them (cf. imperfect transmission). If one trustworthy information
provider sends information to several persons via the push mode (i.e. multicasting), the imperfect
transmission problem may be avoided and it only remains the synchronization issue.
Finally, if a viewpoint that is consistent to the proposed definitions is adopted, the push mode isn’t
involved by the retrieval issue because the communication is instantiated by the computer.
The pull mode permits to limit the information overload because the user doesn’t have to handle any
disturbing incoming information flow. The information is picked up from a repository when needed.
The pull mode suppresses the well-known issue of ‘just-in-case-they-need-it’ push publishing (Thorpe
and Mead, 2001) (which must be considered, in fact, as a misuse of the push mode). The greatest
advantage of the pull mode is that it doesn’t offer the user an easy way to abusively trigger hundreds
of information flows. There is however a counter part of this situation: the dependence on the initiative
taken by the user. The pull mode isn’t able to resolve a situation where a project participant doesn’t
instantiate any connection to the repository.
The pull mode presents also an advantage in terms of information symmetry. Because one
information source is hypothesized to be up-to-date, it is easy to know where to collect the latest
version of any information piece. As already mentioned, the efficiency of the pull mode in this domain
primarily depends on the refresh rate of the data. More precisely, in most of the pull mode
implementations, a temporary buffer storing previously retrieved information is used to optimize the
performance. Due to this feature, the data may be accessed while the synchronization mechanism is
deactivated and the risk of consulting old data reaches the same level as in the multicasting solution.
The information retrieval issue must probably be considered as the major drawback of the pull mode.
As the amount of information to be searched increases, the difficulty to rapidly find the required data
becomes more important. In the worst case, the required effort may discourage the person to search
the repository and consequently to store anything into. Therefore, an efficient pull mode
implementation must provide structured information combined with user-friendly retrieval tools.
THE SITUATION IN THE CONSTRUCTION INDUSTRY
In practice, most of the solutions proposed to manage construction projects combine some push and
pull functionalities. This can be illustrated with the example of a project specific web site (PSWS),
which is nowadays probably the most promoted approach. While this solution is often presented as
the perfect example of pull mode, it includes, most of the time, some push-based functionalities (e.g.
e-mail technology used for data update notification or invitation to meetings). The push-oriented
approaches also often include some pull functions. For instance, an electronic message sent to invite
the receiver to visit a website may be qualified in this way. The effective management of a project
requires thus an environment able to deal with push and pull communications.
Instead of improving pull-based project management environment with some push functionalities, we
have chosen to adopt the symmetric way: designing an enhanced push-based tool – the BBeLink2
prototype – including some pull features. This approach is based on the observation that web sites
present widely discussed advantages but also some drawbacks pointed out by quite few authors.
Most of the web site systems face the same problems than those encountered for years by integrated
project databases (IPDB). For instance, information retrieval becomes more difficult as the amount of
data increases, data partitioning and access rights may be very difficult to manage (see Amor & Faraj,
2001, for a comprehensive discussion on misconceptions about IPDB). Moreover, they cause the
emergence of new kinds of risks for the companies, especially for the smallest ones:
• dependence on the (potentially rival) company hosting the web site,
• dependence on the usage by all projects participants,
• confidentiality of the data,
• risk of failure of the web site operator.
Finally, the greatest challenge for the PSWS is probably to bring the construction professionals to
replace some of the traditional point-to-point communications by the consultation, relying on personal
initiative, of a central repository. The effective use of a PSWS requires the project participants to
change their way of performing some of their tasks. Unfortunately, some recent empirical experiences
have concluded that ‘even six years is a short time to change human attitudes to the means of
communication used, in spite of impressive new technologies being available.’ (Howard and Petersen,
2001) In conclusion, ‘the organisational issues surrounding the use as well as the psychology
involved in getting all participants in projects to accept using new technology are now in focus.’ (Björk,
2002)
From this perspective, the BBeLink2 approach is less ambitious. It is based on the strategy ‘which
does not present technology as a new mode of organisation but only as the way to introduce its
necessity’ and this approach ‘is not unfounded and has been successfully used within other industries
[than the construction]’. (Brousseau and Rallet, 1993). This approach also takes into account the
importance of proposing technologies that may be rapidly accepted, which has been recognized as a
fundamental element of acceptation for the small companies (Mundim and Bremer, 2000).
BBeLink2 has been designed to handle both point-to-point communications like classic e-mails and
multicasting dissemination, without using a central web site. The multicasting diffusion of information
may be instantiated only by the project creator. It is used to diffuse global project properties that
present some interest for every participant (e.g. picture of the current state of the building site, project
participant list, building description). This feature is intended to increase the confidence of every
participant in the data validity. A technological description is provided in the next section.
BBELINK2 TOOL
Technology description
The BBeLink2 prototype is based on the creation of a community. Every member of the system uses
a client application that handles the connections to a server. There are only two centralized modules
shared by all the participants: a communication server used to transfer the messages and a central
repository containing the identification of the people/companies allowed to use the system. There isn’t
any central repository for project data.
BBeLink2 includes two major types of functionalities: communication modules and storage modules.
The communication issues are handled by implementing two kinds of messages that may be
transferred among the client applications: user messages and system messages. The former may be
compared to classic e-mails that would be enhanced by some value-added features (e.g. reliable time
synchronization for message dating, inclusion of meta data describing some message properties).
They are dedicated to be composed and read by human beings. The user messages include some
primary and secondary data. The primary data are those that should be read by the receiver in all
cases (i.e. sender, sent date, subject, content text). The secondary data regroups those that may be
read by the user if he wants to process the message in a comprehensive manner (e.g. for sent
messages: reception status for every receiver; for received messages: attached files). The system
messages are exchanged between computers for performing some low-level tasks in a transparent
manner for the user (e.g. acknowledgement mechanism for every user message, synchronization of
global project data via multicasting).
The storage modules are dedicated to store the information related to the projects. This data
describes two kinds of objects: the communication instances and the project properties. The former
concerns the description of all the information flows (IFs) occurring among the project stakeholders,
independently of the communication media used. BBeLink2 includes a repository for storing all IFs. It
distinguishes a structured and an unstructured part in an IF description. The structured part is
implemented as a set of standardized meta-data to assign to any IF. Some are very usual, such as
the identification of the sender, and others are more original, such as the content type (technological,
administrative, commercial…) or the communication media used (phone, fax, meeting, e-mail…). The
meta-data set has been instantiated from a three-step process. First, a literature review has provided
a list of potential values (e.g. list of project stages). Second, a limited list of values has been built,
which appeared as an acceptable compromise between usability and utility. Third, the resulting list
was revised by a partner enterprise active for several years in the AEC industry. The unstructured
part allows to describe the IF with free text and to attach any kind of file (picture, drawing, contract…).
The IFs linked to the BBeLink2 user messages form a specific subset. Because the user messages
are instantiated inside the BBeLink2 environment, the corresponding IFs are automatically created in
the IFs repository. The latter concerns the project itself and is implemented by project properties (e.g.
building description, construction site picture, participants list). BBeLink2 distinguishes global and
local project properties. The former are defined by the project creator. They are diffused to the project
participants and synchronized by multicasting. The latter are also synchronized but they may be
modified by the project participants.
BBeLink2 and the information management issues
The BBeLink2 concept must then be confronted with the three basic issues related to information
management that have been previously mentioned. In order to facilitate the understanding of the
different discussed elements, the Figure 3 provides a global view of the BBeLink2 features, in terms
of information dissemination modes.
The first issue to analyze is the information overload. BBeLink2 includes several features intended to
handle this issue.
• The server handles only messages of the members of the community. Moreover, it may be
configured to block the messages of any member showing an inappropriate behaviour. This
mechanism may thus limit the risk of information overload due to spam.
• The BBeLink2 clients integrate a filtering mechanism. A filter is a logic expression defined
using primary (preferably) or secondary data of the user messages. A filter may also be a
combination of existing filters. The filters are evaluated for each incoming message. If the
expression is true, a specified action (e.g. forward, save in specified directory…) is
automatically triggered. This mechanism may limit the amount of incoming flows to handle.
• The distinction between primary and secondary data in the user messages also limits the
information to handle during the first step of the cognitive processing. Moreover, the user may
choose which (primary) meta data of the messages should be displayed in the graphical
interface.
• It is possible to generate summary reports about the information flows that have occurred
during the project. This may help the professionals to rapidly get an overview of the current
state of a project.
• The user may deactivate the push-based warning messages sent when some global project
properties are updated.
• The user interfaces used to handle user messages and information flows are very similar, in
order to reduce the learning effort for the user.
Server
{System Messages, User Messages}
User A
Application client A
Project properties
Messages A
Information
flows
A
Global
User Messages A
System Messages A
Local
Primary
data
Secondary
data
Push Pull Pull Pull
Push
(update warning)
C
o
m
m
u
n
i
c
a
t
i
o
n
m
o
d
u
l
e
s
S
t
o
r
a
g
e
m
o
d
u
l
e
s
Figure 3 BBeLink2 conceptual model (push-pull viewpoint)
The second issue concerns the retrieval of information. The design of BBeLink2 has tried to structure
the data as far as possible, in order to ease the information access.
• The whole BBeLink2 environment is organized by project. Every message and every
information flow is linked to a specified project.
• Every user message and information flow is characterized by a (similar) set of meta data. This
allows an efficient sorting and retrieval of any instance of communication.
• The filtering mechanism may be used to automatically organize the incoming user messages
in a given way, specific to the company.
The information asymmetry is the last issue to be discussed.
• The multicasting system guarantees that the global project properties are identical for every
project participant. Indeed, only the project creator is granted to set and modify them. The
local replication that is stored on the computer of each participant is encrypted and may not
be accessed in write mode.
• The information synchronization depends on the connection frequency to the Internet. As
soon as a new connection is instantiated, every new modification of the global project data is
automatically updated on the local instance. This update is transparent to the user and can’t
be deactivated.
• BBeLink2 also provides a reliable dating system. Each time value set in relationship to a
message is collected from an external time server. In conjunction with the low level,
permanently activated acknowledgment mechanism, this permits to be really informed on the
reception status of any message, for any receiver. The sender may thus also be aware of a
lack of synchronization.
• By integrating a digital signature implementation, BBeLink2 prevents the corruption of the
transferred data.
CONCLUSIONS
In many discussions, the pull and push mode are presented as rival modes of information
dissemination while they are in fact complementary. Moreover, the technological viewpoint often
dominates the functional perspective. This paper describes a proof-of-concept tool, called BBeLink2,
which mixes push and pull functionalities. It is original in many ways but the absence of a central web
site is certainly one of the most worthy to be mentioned. The next step aims to get a sample of
architects to evaluate the tool, especially in terms of utility, usability and intend of use.
REFERENCES
Acharya B., M. Franklin and S. Zdonik (1997), Balancing Push and Pull for Data Broadcast in
Proceedings of ACM SIGMOD Conference, May 1997, Tuscon, Arizona.
Amor R. and Faraj I. (2001), Misconceptions about integrated project databases, in Electronic Journal
of Information Technology in Construction,http://www.itcon.org/2001/5/, accessed 12 Dec.
2002.
Basiel A. (1999), Push & Pull technologies using a WIDE approach – Web-based Instructional
Designed Environment, 3rd Human Centred Technologies Postgraduate Workshop 1999,
Brighton.
Björk, B.-C. (2002), The impact of Electronic Document Management on Construction Information
Management, in Proceedings of the CIB w78 Conference, Distributing knowledge in building,
June 2002, Aarhus, Denmark.
Brousseau E. and A. Rallet (1993), Développement des systèmes télématiques et évolution des
relations interentreprises dans la construction, Rapport d’une étude financée par le Plan
Construction et le PIRTTEM, Université Paris-Sud, available on demand fromhttp://www.jm.u-
psud.fr
Cerami E. (1998), Delivering push, Mc Graw Hill, NY.
Clark A. M., B. L. Atkin, M. P. Betts and D. A. Smith (1999), Benchmarking the use of IT to support
supplier management in construction, in Electronic Journal of Information Technology in
Construction,http://www.itcon.org/1999/1/, accessed 12 Dec. 2002.
Farhoomand A. F. and D. H. Drury (2002), Managerial information overload in Communications of the
ACM, October 2002, Vol. 45, No. 10, pp. 127-131.
Howard R. and E. Petersen (2001), Monitoring communications in partnering projects in Electronic
Journal of Information Technology in Construction,http://itcon.org/2001/1, accessed 26 Nov.
2002.
I.A.I. (2002),http://cig.bre.co.uk/iai_uk/iai/page2.htm, accessed 11 Dec. 2002.
Koskela L. and R. Vrijhoef (2000), The prevalent theory of construction is a hindrance for innovation,
in Proceedings of International Group for Lean Construction 8th Annual Conference,http://www.susx.ac.uk/spru/imichair/news/template.cfm?content=iglc-8.cfm, accessed 14 Dec.
2002.
Mundim A. and C. F. Bremer (2000), Design of a computer-supported cooperative environment for
small and medium enterprises, 2
nd
IFIP Conference: Managing Business-to-Business
Cooperation, Dec. 2000, Florianopolis, Brazil.
O’Conner J.T. and R. L. Tucker (1986), Improving Industrial Project Constructability, in Journal of
Construction Engineering and Management, Vol. 112, No1, March, pp 69-82.
SearchWebServices.com (2002),http://searchwebservices.techtarget.com/, accessed 14 Dec. 2002.
Thorpe T. and S. Mead (2001), Project-Specific Web Sites: Friend or Foe? in Journal of Construction
Engineering and Management, September/October 2001.
Whatis.com (2002),http://whatis.techtarget.com/, accessed 14 Dec. 2002.
doc_460838595.pdf