<< 1 >>
Rating: 
Summary: What you need to consider when choosing a modeling technique
Review: OM&UID is a book about developing new object-oriented methodologies for interactive software. Nine different methods are presented by an international team of software experts like Larry Constantine and Philippe Kruchten. All the authors are trying to solve constraints or deficiencies in existing methods. Since these are all new or experimental techniques, each author explains exactly what problem s/he is trying to solve, where the new method might be best used, and how it worked in practice. Most of the sections work through a couple of cases, so you can see how the method works. A couple of the writers have pointed out how difficult current heavy-weight methodologies are to use. The models generated, unless the modeler is extremely experienced, are usually not correct. What's more, as the first chapter notes, the modelers don't realize that their models are bad. A couple of writers have tried to deal with the problem that business customers can't understand UML-style notation, and don't mentally describe their jobs in terms of classes or windows. That cuts customers out of the system design process at exactly the point where they should be most engaged. The editor repeats what is generally recognized: that very few people use a methodology as such. Most of us use a grab bag of techniques from a mix of methods, heavily customized to our own needs. Mark van Harmelen's book may be best addressed to those who use mixed methods, because it helps us to see how experienced architects decide which techniques to use in different circumstances and how we can determine whether we were successful.
Rating:
Summary: What you need to consider when choosing a modeling technique
Review: OM&UID is a book about developing new object-oriented methodologies for interactive software. Nine different methods are presented by an international team of software experts like Larry Constantine and Philippe Kruchten. All the authors are trying to solve constraints or deficiencies in existing methods. Since these are all new or experimental techniques, each author explains exactly what problem s/he is trying to solve, where the new method might be best used, and how it worked in practice. Most of the sections work through a couple of cases, so you can see how the method works. A couple of the writers have pointed out how difficult current heavy-weight methodologies are to use. The models generated, unless the modeler is extremely experienced, are usually not correct. What's more, as the first chapter notes, the modelers don't realize that their models are bad. A couple of writers have tried to deal with the problem that business customers can't understand UML-style notation, and don't mentally describe their jobs in terms of classes or windows. That cuts customers out of the system design process at exactly the point where they should be most engaged. The editor repeats what is generally recognized: that very few people use a methodology as such. Most of us use a grab bag of techniques from a mix of methods, heavily customized to our own needs. Mark van Harmelen's book may be best addressed to those who use mixed methods, because it helps us to see how experienced architects decide which techniques to use in different circumstances and how we can determine whether we were successful.
<< 1 >>
| 
