I wonder if anyone can assist me, I am trying to do a use case diagram from the activity diagram that I drew but I am not very good in use case diagram; can anyone guide me if what I did here is correct please.
You can see the diagrams at this link:
http://imageshack.us/photo/my-images/857/activitydiagram.png/
http://imageshack.us/photo/my-images/816/usecase.png/
Thanks in advance.
Some background first.
Activity & Use Case Diagrams usually depict behaviour. Activity Diagrams usually describe a process (or a linked sequence of activities). Yours doesn't really do that. That's OK - as long as the diagram works for you that's fine.
It's also normal to design an Activity Diagram as the process to realise a Use Case. You're asking to go the other way. Again, not conventional - but not impossible.
Use Cases are normally written as verb statements and describe actions a User would want to perform (e.g. "purchase book", "checkout"). Yours are nouns.
OK - so to your question.
Identifying the Use Cases involves asking the question: "What can Users do?". Looking at your Activity Diagram, I'd guess the Use Cases would be something like:
Take a Picture [with Camera]
View Photos
View Map
View Current Location on Map
Each of those would be a Use Case. There are probably others (what can a User do with Favourites?) but that should give you the idea.
hth.
Related
With this Use Case diagram, I'm unsure if I'm currently doing it correctly or not. Checked online and each site is giving me a different answer. I just need a direction or someone to correct me.
This is interesting as I am currently going through the exact same discussion with a client...
I would say "no" simply because I am unsure of what this actually represents.
A use case must bring a benefit to a user, and I really don't this in this diagram. What does Actor0 actually trying to accomplish? I do not get this from the diagram.
Did you write down statements before trying to do this? I find that going to a whiteboard and writing down what your actor's goal and how you see this happening so that the actor achieves the goal is a good start to better understanding the use case and differentiate the system being built from external systems. One way to do this is through a set of user stories (or use case slices).
And answering Thomas Kilian's excellent questions will also help you in this task!
Also, you are doing something that I see too often from people starting with use cases (and sometimes with UML experience): use case decomposition as a set of actions - basically trying to model a set of activities by breaking it down. Doing that is typically the second step in your analysis (e.g., by using activity and sequence diagrams).
No. Only "Register" (what?) and "Add/View Details" (which?) can be regarded as use cases since they add value. Neither "Login" nor "Main Menu" are use cases. The first is a constraint and the second an implementation detail.
This picture is a capture of my activity diagramm of the use cases consult:
The diagram is fine so far. However, you should add meaningful guards to the transitions coming from the decision node (though the activity names might already make that clear, but my French isn't good enough.
I came across a situation where a use case has two options to choose from.
Example: Use case "Process data" has options like "Approve" and "Decline"
If approve, update the inventory.
If decline, file the information.
I know how to represent this in an Activity Diagram but I have no idea of showing it in Use case diagrams.
There are two choices here.
A) You make two use cases instead of one. Approve Data and Decline Data
B) You make one use case Process data. In the main scenario you describe the steps for Approve Data and in the exception or alternate scenario you describe the steps for Decline Data. In this case the difference is not visible on the use case diagram, but that is OK. Use case diagrams are meant to give you an overview of the big chunks of behavior in your application, not of the details.
The choice for alternate or exception scenario depends on your post-condition(s). If the Declinescenario meets the post-condition then it is an alternate, else it is an exception scenario.
The term Data is not very specific. I would prefer something more meaningful then data like Order, Application,...
I can't know for sure without knowing all the details, but option B) is probably the best.
A usecase has a specific goal for a specific actor.
Identify the Actor for the usecase(s).
Identify the Goal of the Actor.
Define the usecase with a strong specific verb, process is weak unspecific verb.
I find the requirements ambiguous, it seems you may be conflating two separate Actors.
See the Uml Style Guide : Usecase
You create two views of the same usecase. In the first view you create Process data with Approve and a second one with Decline. Each view is a different usecase diagram but using the same usecase.
It is important to make a clear differentiation between graphical views and model view. I mean that graphical usecase diagrams should be a view which is extracted from the UML model.
if not then use this link http://msdn.microsoft.com/en-us/library/dd409427.aspx
I have to develop a CRUD application, that will be coded in php.
I have 3 main actors (Users, Administrators and Doctors -- this is for an hypothetical
hospital), each one with different Use Cases already defined.
Although I feel the Use Cases are more than enough to successfuly model the Class Diagram, I am being specifically asked to also include DataFlow Diagrams into the project's documentation.
I've been reading about DataFlow diagrams, and it seems you usually have first of all a Level 0 DataFlow Diagram, to which they call Context Diagram.
Being that this is basically a 3-tier application with 3 different Actors, how should I model the Context Diagram?
Being that a Context Diagram is supposed to just tell us what comes in and what comes out of our System, I can't imagine anything more interesting/descriptive than the following diagram:
Is this supposed to be something like this, or am I totally missing the point? This php page will connect to an Oracle database, but I guess that if the idea is to consider the System as a whole in the Context Diagram, I should "hide" that fact in the above diagram.
Where should I go on from here? I know I should "zoom" the System process to something more detailed. Maybe the next step would be to depict each one of the User Cases in a DataFlow diagram? Do I include repositories of data, already? For example, one for Users, other for Doctors and yet another for Administrators?
Thanks
Are you sure there's nothing else the system interacts with? e.g. diagnostics input, etc.?
If not then your context diag is basically ok - although I'd probably show each entity once and use double headed arrows. I'd agree with your reasoning for the db - it's part of the system, not external to it - so don't show it on the CD.
As for next steps, again you're on the right lines. Try modelling the flow for each Use Case as a DFD. DFDs are very useful for illustrating processing-intensive apps. Difficult to know if that's a good match to your problem or not.
You'll find DFDs are also useful for driving out and validating your Class Diagram. In fact, that's one of their strengths: datastores on the DFD should correlate with the contents of your class diagram (not necessarily one datastore to one class though). So do include datastores as you work through the processes. You'll find it drives out more than just the Actors.
hth.
Some remarks:
You DFD does not tell me much, except that Users, Administrators and Doctors use it, but it gives me no clue what they get from the system (except "Output Data"). IOW the context diagram does not give me the slightest idea, what the system does.
Admittedly, if the system is large, then it can be difficult to describe the dataflows in few words, but just about anything is better than "data".
The fact that the system is a 3tier architecture is irrelevant for the DFD. This is an implementation detail. DFDs are an analysis tool. You describe what you want the system to do, not how this is achieved.
I find it particularly useful to focus on the outgoing flows. While Users, Administrators and Doctors provide input to the system, this is most likely nothing they want to do. It is something they have to do in order to get the desired output.
On a Use Case diagram can you show things that an actor cannot do, for example because they won't have permissions to do it?
Or is it just implied due to the fact that they won't have a line joining them to the particular use case?
If the Use Case you are diagramming is the case where an actor attempts to do something that is not permitted and is then denied, then yes, I would show it.
Otherwise, I would stick to only including things that are actually part of the use case.
No. An Actor would be connected to everything that he can do. If the Actor can't do it, then it's not shown.
This is what alternate paths are for. The basic path (a.k.a. happy path) will show what happens when the correct Actor initiates the Use Case. In the alternate paths you can show what happens if the wrong Actor attempts to initiate it.
You might model Role actors that can do the task. You could then have another use case that has the original actor attempting to acquire the given Role.
IMHO this question and most of the answers give a wrong impression about the way use cases should be used.
Use case was intended as a requirements technique that uses natural language. It is most and quite effective that way.
It can be a thoroughly destructive technique when it is combined with too much UML/modeling. Structured modeling of use case texts for example by modeling main and alternative flows using UML Activity Diagrams is a tried and tested way for example to create Use Cases of Mass Destruction.
A use case diagram can be useful but we should remember the purpose of use case as a technique which is first and foremost to identify the user goals a system should support. Subsequently we can capture more details using natural language in use case texts using main flow, alternative flows etc.
Using diagramming tools we can visualize some simple information:
- For each user goal we can create model element type Use Case.
- Show system boundary using a box for the system with use case elements in it.
- Create a relationship between actor and use case to show the actor has a particular goal against the system.
Keeping an up-to-date list of actors mapped to goals is however of secondary importance. Doing a stakeholder analysis, drawing up lists of actors is a means to identify the users goals. After user goals have been identified it is strictly speaking not longer necessary to keep the lists of actors around.
If we are asking questions about how to put user permissions in a use case model we are most likely trying to capture too much information. We should abstract model elements away so that the model does not try to answer/capture these type of detailed design questions.