Advantages of Dimensional Modeling

Hi
I have few questions on advantages of dimensional modeling as provided in the book DWH Lifecycle Toolkit by Ralph.
(1) In the book , on Page 150, it is mentioned that "Aggregation software packages and utilities depend on a very specific single structure of fact and dimension tables that are dependent on dimensional models".
I would like to know How BI Tools (like BusinessObjects) identify Dimensions and Facts as when the datamodel is implemented in RDBMS like Oracle, there is no special tag which identifies a table whether it is a dimension or fact?
(2) On Page 148, it is mentioned that the third strength of DM is the model is gracefully extensible. Some typical reasons mentioned are like SQL ALTER TABLE command, Adding un anticipated facts, new dimensions etc.
I think most of these strengths are equally applicable to a model built in a non dimensional approach also. Can these be considered as exclusive strengths of DM ?

Please note that these questions are being posted only from the perspective of gaining good understanding and improving knowledge in DM front

Thanks in advance

1. You define the type of table in the BI tool.
2. If the model is normalized then yes.

1. From a BOBJ point of view, you do, in fact, identify objects as being dimensions or measures. So there is an implied understanding of dimensional structure. Also, when you define contexts in the Universe you are essentially bounding star schema. Sure you can do the same with normalized tables, but it is much more difficult. When you query aross contexts, BOBJ does the standard dimensional process of querying each context individually, aggregating it and joining on common dimensions. Also, don't ignore the aggregate awareness feature in BOBJ.

2. A normalized model is more difficult to change under certain conditions. If the change involves adding new data, that, based on normalization rules, requires you to create a new entity, you need to create the entity as well as add foreign keys to which ever other entities are affected. Whereas, in a dimensional model, if you need to add some attributes to a dimension, you do just that. Of course, if you need to add a new dimension, you are faced with similar challenges as adding an entity in a normalized model. The difference here is that the decision to create a new dimension is not forced on you by a prescribed form.