Career Forums

 Hi

I am a BA and I am writing use cases for a very large product reverse engineering project. We have a VB based application which has many batch processiing capabilities. I am writing for the use case for one of the complex batch system which does not have any screen. There is not going to be any actor interaction in this use case. First of all, for a batch system do we elicit the requirements in a Use case doc  or any other document? This is my approach.. I have broken the entire batch system into logical chunks where each chunk represent a unique functionality and I am writing the use case for these chunks. Since there is sequential flow, I am mentioning the Post Condition of one use case will become a pre condition of another. Is this approach correct? I have so far written use cases where there is actor-system interaction and this is the first time I am attempting to write for a batch system. Can someone suggest how do I write the use case. if you have any specific template to acheive my requirements, pls do post it..

Thanks in advance.

Chandra

Chandra,

1: Listen to Kimbo!

2: What you are doing is reverse engineer batch processes.  You are going from a Platform Specific Model(PSM: physical Model) to a Platform Independent Model (PIM: logical model). Have a brief read about model driven architectures (MDAs at the OMG site), just to get an idea.

Here is what I'd do if I was a Techo/Solution Architect (not a BA). 

• I'd take the VB/sql code (assuming that you do have access to these things) and reverse engineer it within a inexpensive system like SPARX Enterprise Architect. See http://www.sparxsystems.com/uml_tool_guide/code_engineering/reverseengineersourcecode.html. This should give me a list of objects, which I'll then group into a component diagram (UML2) for the Physical View.
• I'd also create a high-level  component diagram (UML2)  and their inter-relationships for the logical view (I guess this is what you are trying to achieve). Note, this is not going to be definitive but at least you will have the logical and physical components and their inter-relationships. SPARX allows you to create these inter-relationships. It might be a good idea to group the logical components into a hierarchy that includes model (those components that update the database),view (components receiving/sending messages between other components) and controller (those components doing the logic) components
• I'd take a function like say "month end invoice run" (a batch process) which has a "time clock/calendar" as an actor, and create a sequence diagram with the logical components. 

I'd suggest you collaborate with your solution architects to create a repository of your artefacts; again I'm assuming that at some point you'd like show functions, data and user interfaces etc and their inter-relationships.

All the best!

warm regards,

K