essential features of the real system it attempts to portray. We can increase our
confidence in the model further by comparing its results to other models for the
same or similar systems. The latter approach is of particular interest if the model
was descriptive rather than predictive. Finally, we hope that we develop, through
practice in modeling, experience and new understanding of system dynamics that
enable us to more easily detect problems in model specifications. This is a learning
process and we hope our book makes a significant contribution.
1.9 Extending the Modeling Approach
The models developed in this book are all built with the graphical programming
language STELLA. In contrast to the majority of computer languages available
today, STELLA enables you to spend the majority of your time and effort on
understanding and investigating the features of a dynamic system, rather than
writing a program that must follow some complicated, unintuitive syntax. With
its easy-to-learn and easy-to-use approach to modeling, STELLA provides us with a
number of features that enhance the development of modeling skills and collaboration of modelers. First there is the knowledge-capturing aspect of STELLA. The
way we employ STELLA leads not only to a program that captures essential
features of a system, but also more importantly results in a process that involves
the assembly of experts who contribute their specialized knowledge to a model of a
system. Experts can see the way in which their knowledge is incorporated into the
model because they can pick up the fundamental aspects of STELLA very quickly.
They can see how their particular part of the whole is performing and judge what
changes may be needed. They can see how their part of the model interacts with
others and how other specialists formulate their own contribution. As a result, these
experts are more likely confident that the whole model is able and accurate than if
they had entrusted a programmer with their insight into the system’s workings.
Once engaged in this modeling process, experts will sing its praises to other
scientists. This is the knowledge capturing aspect of the modeling process.
These same experts are likely to take STELLA into the depths of their own
discipline. But most important, they will return to their original models and repair
them to meet broader challenges than first intended. Thus models grow along with
the expertise and understanding of the experts. This is the expert-capturing part of
our process. It is not only based on informed consensus but it has the possibility of
continuing growth of the central model.
So we avoid the cult of the central modeler on the mainframe computer, who
claims to have understood and captured the meaning of the experts. We avoid
producing a group of scientists who, unsure how well their knowledge has been
captured, and so, in their conservative default position, deny any utility or even
connection to the model.
The process is not risk free and it requires delicate organization. The scientists
who do participate in the modeling endeavor will reveal an approach to their field.
26
1 Modeling Dynamic Biological Systems
confidence in the model further by comparing its results to other models for the
same or similar systems. The latter approach is of particular interest if the model
was descriptive rather than predictive. Finally, we hope that we develop, through
practice in modeling, experience and new understanding of system dynamics that
enable us to more easily detect problems in model specifications. This is a learning
process and we hope our book makes a significant contribution.
1.9 Extending the Modeling Approach
The models developed in this book are all built with the graphical programming
language STELLA. In contrast to the majority of computer languages available
today, STELLA enables you to spend the majority of your time and effort on
understanding and investigating the features of a dynamic system, rather than
writing a program that must follow some complicated, unintuitive syntax. With
its easy-to-learn and easy-to-use approach to modeling, STELLA provides us with a
number of features that enhance the development of modeling skills and collaboration of modelers. First there is the knowledge-capturing aspect of STELLA. The
way we employ STELLA leads not only to a program that captures essential
features of a system, but also more importantly results in a process that involves
the assembly of experts who contribute their specialized knowledge to a model of a
system. Experts can see the way in which their knowledge is incorporated into the
model because they can pick up the fundamental aspects of STELLA very quickly.
They can see how their particular part of the whole is performing and judge what
changes may be needed. They can see how their part of the model interacts with
others and how other specialists formulate their own contribution. As a result, these
experts are more likely confident that the whole model is able and accurate than if
they had entrusted a programmer with their insight into the system’s workings.
Once engaged in this modeling process, experts will sing its praises to other
scientists. This is the knowledge capturing aspect of the modeling process.
These same experts are likely to take STELLA into the depths of their own
discipline. But most important, they will return to their original models and repair
them to meet broader challenges than first intended. Thus models grow along with
the expertise and understanding of the experts. This is the expert-capturing part of
our process. It is not only based on informed consensus but it has the possibility of
continuing growth of the central model.
So we avoid the cult of the central modeler on the mainframe computer, who
claims to have understood and captured the meaning of the experts. We avoid
producing a group of scientists who, unsure how well their knowledge has been
captured, and so, in their conservative default position, deny any utility or even
connection to the model.
The process is not risk free and it requires delicate organization. The scientists
who do participate in the modeling endeavor will reveal an approach to their field.
26
1 Modeling Dynamic Biological Systems