2

I'm learning how to design a Data Flow Diagram (DFD) for a system. I have a question that I can't answer myself. Let me give a scenario:-

In a system, a Trainer wanted to view his/her own profile information and a Trainee also want to view his/her own profile information.

Can the process (Retrieve profile information) be used by both entities:-

DFD 2 Entities 1 Process

Or must be separated as they aren't viewing the same data:-

DFD 2 Entities 2 Processes

So, which one is right?

Christophe
  • 68,716
  • 7
  • 72
  • 138
Kaydarin
  • 291
  • 1
  • 9
  • 23

3 Answers3

1

The issue in this question is that the data store and the entities are referred to as if they were specific instances:

  • The Trainer A views profile in the Trainer A file
  • The Trainee B views profile in the Trainee B file
  • But what happens with Trainer C? Would he/she also see Trainer A profile? Or just Trainer C profile?

As a first step you need to clarify the data stores: if each trainer has his/her own file, call the data store Individual trainer file to clarify that there is a one to one mapping. If these files are stored centrally somewhere or could be accessed by more than one person, call the data store Trainer profile store or Trainer repository to clarify that the process has access to many files at the same time (and will need to select the right one). Finally, if trainer and trainee profile contain the same data, but it's just that they are owned by different persons, merge the two data stores (e.g. like if it's a database table combining all the profiles).

Another issue is that the narrative only presents a specific scenario: at a given moment int time, Trainer wants to see trainer profile, and Trainee wants to see trainee profile. But what shall the process allow in general? Should a Trainer be able to view only trainer profile, or also trainee profiles if he/she'd wanted? An only his/her trainees or all the trainees?

To draw the right diagram you have to generalize your scenarios and entities and provide the answers above.

Depending on the result:

  • if the process only allows a category of entities to view some kind of data, go for option 2.
  • if the process allows some entities to consult several kind of data, go for option 1. The process specification need then to clarify who has access to what.
  • if the profile data are of the same kind, and it's just the owner who changes , then go for option 3: one data store, one process, but different entities, the process specification clarifying what entity can access to what profiles.
Christophe
  • 68,716
  • 7
  • 72
  • 138
0

Kaydarin,

From my understanding the key for your definition is in the word "and". You can also see that trainee and trainer are not viewing the same data...

If process involves multiple entities providing different pieces of information to be processed for a common output, it is more likely that a single process would apply.

Enjoy designing,

Andres

Andrés Lescano
  • 1
  • 1
0

The second model one is more flexible since the processes have a higher cohesion and could be better reused since coupling is lower.

It is advisable to not only look at the functionality of an process but also look at reusability and maintenance. In the second solution you do not have to test retrieve Trainer profile process if the data structure of a Trainee object changes.

martin bruijn
  • 1