All,

Some while ago I started a new assignment at an Electricity Distribution firm. The first thing I did was to create a logical data-model (domain model) to get an understanding of the key entities in the industry.

I like the MVC story on http://www.modernanalyst.com/Careers/InterviewQuestions/tabid/128/articleType/ArticleView/articleId/953/What-is-ModelViewController.aspx

It shows that the controller and view are dependent on the data model; therefore its always better to first model the data then pursue the controller and view aspects. OK, it is design architecture; but how many, during elicitation, ask questions from a data driven approach?

 Is starting from a data perspective first a good strategy for an understanding and eliciting requirements?

How do you approach new assignments; in particular, what techniques do you use to fast track your understanding of the business?

warm regards,

K

Hi:

Data models are relatively inert - they do adequately prod the analyst through discovery.  Functional modeling is much more rigorous in prodding discovery.   To fast track, do functional modeling first and then use such as a primary input to data modeling.

To fast track functional modeling, think, "What are the most essential requirements?"   Typically, as a function is defined, at least at higher levels of abstraction, by inputs and outputs, you want to employ a modeling techniques that makes areas where you need to conduct more discovery of inputs and outputs glaringly obvious.

Tony

K:

To "get the ball rolling", doing some data modeling first seems OK.   But, my main point is that data modeling is not going to prod the analyst through rigrous discovery.   I know of many analysts who think "why functional model, I will just data model".  Such an approach derails discovery.

I also firmly believe that analysis should start at as high a level as possible.  Else, there is a tendancy for the analyst to, as DeMarco used to say,  "drown in an ocean of detail".

Tony