When using the Rational software development tools to perform requirements management there is a potential source of confusion as to what tool to use when and a perceived overlap of functionality. This is more so when using a use-case orientated approach (as opposed to a traditional prose declarative document approach) to specifying functional requirements. By what method should use-cases be documented and how should this information be represented?
Having looked at this and used these tools recently the following would be my recommendations as to how to effectively use these tools in a use-case orientated project.
RequisitePro is a requirements management tool, it is not a general document management tool and is therefore generally used to stored only those artifacts which are products of the ‘Requirements’ discipline. Rational Software Architect (RSA) is a modelling tool but can be used to manage a use-case model and this model can have external non-modelling artifacts associated (and linked) with it.
Given that the requirements being managed can be broken down into
- Stakeholder Requests
- Features
- Software Requirements
and that software requirements themselves can be further broken down (in a use-case based project) into
- Functional (Use-Cases and Supplementary)
- Non-Functional (Supplementary)
Then the overlap in this context is where is the use-case based information managed. The use-case information tends to manifest itself as use-case diagrams (model elements) and use-case specifications (documents).
Based on my experience the best way to manage this information is for the textual requirements information (e.g. Stakeholder Requests, Features and Software Requirements – including any Use Case Specifications) to be managed within RequisitePro and this can therefore be used to track changes and establish traceability between elements. The model disgrams and elements (including actors, use-cases and relationships) are best managed within RSA.
With the above the understanding the question becomes one of, how best to link these to ease navigation and establish traceability. Assuming that the stakeholder requests and features have already been captured within RequisitePro and a project respository exists then this can be linked to from withi RSA. Following a UML model and associated project having been created within RSA open the ‘Requirement Explorer’
- Window->Show View->Requirement Explorer
This will open a new window within the RSA framework, then select the button entitled ‘Open a RequisitePro Project’ and browse to the relevant *.rqs file associated with the RequisitePro project
e.g. \\{path to project share}\TestProject\TestProject.rqs
If security has been enabled on the project this will require you to login using a username/password. Once the project is open you can link requirements to model elements by dragging the requirement from the ‘Requirements Explorer’ to the desired model element within the ‘Model Explorer’.
Once this traceability has been established then you can select either element within the respective view and trace to the corresponding element from the context-sensitive menu. Equally the details are shown in the ‘Requirements Trace’ view.
