The paper talks about improvements and flaws of the UML version 2 developed by OMG.. He
describes both advantages and shortcomings with examples. The paper states that UML 2 is large
and complex, which is a problem for designers. UML 2 has an option for separating concerns
which makes it easy from the implementation point.

Click here to place an order  for a similar  paper  and have exceptional work done by our team and get A+results

The author says that the second version does
not have a pre-defined set of semantics which can lead to misinterpretation and inconsistency when
the programmer starts defining semantics himself. The author gives a good example of how
mapping implementations can be challenging in UML2. The sequence and interaction diagrams
are improvised over the previous version providing better insight. In model development, a major
problem that the author brings to light is the derivations of dependencies in interactions in a
metamodel, this appears to be an easily accomplishable task but is actually tedious, the model
dependencies like attributes, etc. in a metamodel are pivotal which are scattered throughout the
metamodel. The author suggests aspect oriented modelling techniques and query tools to overcome
the above stated problem. The future scope of research according to the paper is developing tools
and management techniques to overcome the “pitfalls”.
I personally think UML2 is better structured and more expressive than UML. UML 2 has structured
and behavioral models. I agree with the author when he says the designer has to have sufficient
knowledge of previous UML versions to be able to use UML2 because it is built on the same base.
I like the use of “action semantics” in UML2, which makes determining program behavior and
data flow concepts easy. From the examples, I could visualize the concepts of “lifeline” and
“instances”, my understanding is that lifeline represents a participant In any interaction and a class
instance is used in most high level languages. There are other new concepts like “ports” and
“structured classifiers” which are again used to describe decomposition of classifiers. In software
engineering, I think UML2 can be used to obtain a good abstraction of implementation because
most languages use similar concepts. I feel UML 2 is close to reverse engineering, building code
from diagrams, so is UML but this has new concepts.

Click here to place an order  for a similar  paper  and have exceptional work done by our team and get A+results