Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - jps

Pages: [1] 2 3
Suggestions and Requests / Re: FRQ: merge elements
« on: May 14, 2004, 08:09:21 am »
I not only second this, but would love to see support for "Refactoring" in general.

In particular would be the ability to extract a super class from a set of methods in a given class.

Suggestions and Requests / Re: CRC card diagram
« on: April 27, 2004, 03:54:27 pm »
Luca, you can approximate recording the outcome of a CRC session in EA with class diagrams.  

1. Create a class diagram and check the option on it's properties to "Show Requirements."  
2. Create your classes on the diagram
3. On each classes properties select the "Require" tab and enter "Requirements."  These will actually show up on your diagram in a box labeled "responsibilities."
4. Draw associations to the other classes to show the Collaborators

While you could do this to record in EA the outcome of a CRC session, I agree with Tom that you ought to keep EA "miles away" from a CRC session… just stick with the index cards if you are doing CRC.  

If you actually want to do some kind of UML modeling that will get you similar results, consider doing Communication (formerly called Collaboration) diagrams with your team around a whiteboard.  I personally find that to be extremely effective and it is my preferred method for determining both the classes in your system and their methods.

Best regards,


Suggestions and Requests / Re: Priority for use cases
« on: April 16, 2004, 03:15:40 pm »
Hi there.

I agree with Sargasso that from the requirements point of view, the requirements have the priority attribute.

I’m not really sure how either of you are using “requirements” along with use cases.  We take the view that the Use Case Model is the primary requirements artifact.  As Cockburn puts it, use cases "are not all of the requirements - they are only the behavioral requirements but they are all of the behavioral requirements."  Alistair Cockburn, Writing Effective Use Cases, Addison-Wesley, 2001.  In my experience, if I look across the categorization of a set of requirements and do not see that the behavioral, (aka functional) requirements do not make up the vast majority, the team has no real idea what it is they are to implement.

Don’t get me wrong, we absolutely record non-behavioral requirements that make up what RUP calls the Supplementary Spec.  These, however, are not nearly as important as the use cases to our entire software lifecycle.  For us, the use cases drive the estimate, the iteration plans, the architecture (though not entirely), analysis, design, and testing.  

We use Use Cases in a similar way to XP’s User Stories when it comes to planning, which for us makes assigning a priority to use cases very important.  Bruce, you nailed the reason for having a priority on use cases in the last paragraph of your last post.

I can understand that for some customers this may have a low value, but for us (where we have over 100 licenses, and growing) it would be very helpful.

Suggestions and Requests / Re: Priority for use cases
« on: April 14, 2004, 01:47:57 pm »
I actually started down that path but it does not really get me what I want.

I would like to be able to CSV import/export the priority along with the complexity, name, guid, and other items.  It really seems strange that this would not be part of what is (should be) every users primary requirements artifact, the use case.

Suggestions and Requests / Priority for use cases
« on: April 07, 2004, 09:07:57 am »
The documentation states: ‘The priority of this element as compared to other project elements. Only applies to Requirement, Change and Issue types, otherwise ignored. Valid values are: "Low","Medium" and "High"’

It would be highly valuable to us, and I'm sure to the majority of your user community, to have "priority" available for use cases in particular and probably other classifiers in general.


General Board / Re: Parenthesis on a return?
« on: November 09, 2007, 09:40:06 am »
Ok, so in reading the documentation (RTFM! ;)) I see it says that:

"If the Message is a return message, in the Return Value field enter the returned value or type."

This gets rid of the parenthesis, but now I have a colon in front of the return value.  

| :someValue |
|< - - - - - -|

Anyone know how to get rid of that?

General Board / Parenthesis on a return?
« on: November 09, 2007, 09:15:22 am »
On sequence diagrams I'm now seeing parenthesis on a return.

Essentially, I’m seeing something like this:

| message()  |

| response() |
|<- - - - - - |

I used to see this:

| message() |

| response  |
|<- - - - - -|

When creating the return message I’m simply specifying the name of the return value (“response”  in my example above) in the Message text box of the Message Properties dialog box and checking “Is Return” in the Control Flow section.

So, is this now actually per the UML spec?  Or am I not entering the right things in the right places on the Message Properties dialog?



General Board / Re: Bug regarding the way messages are drawn?
« on: September 30, 2006, 07:00:03 am »
Cool, Paolo, thanks for the reference.

If I'm reading the spec correctly then what I'm looking for is "messageident", the return value seems to indicate the actual value that will be returned and placed into the variable specified in "messageident".

From the spec, pg. 512:

<messageident> ::= ([<attribute> ‘=’] <signal-or-operation-name> [‘(‘ [<argument>[‘,’<argument>]* ‘)’] [‘:’ <return-value>]) | ‘*’

I can't seem to find the definition of "messageident" anywhere in the spec... but based on this example on the same page I'm inferring that messageident is what I want.

v=mymsg(16, variab):96  // this is a reply message carrying the return value 96 assigning it to v

I've been able to accomplish this in EA by putting my variable name in the "Assign To" combo box on the Message Properties dialog.  I've also discovered that if I define any instance variables on the class of the sender's lifeline then I can simply select them from the "Assign To" combo box.

In any case, I can't find this (specifically the meaning/use of "Assign To" or "Return Value") documented in the EA documentation.  If its there, would anyone mind pointing me in the right direction?



General Board / Bug regarding the way messages are drawn?
« on: September 29, 2006, 03:07:32 pm »
In build 798 (and in 797) when creating a message between two lifelines, on either a sequence or communication diagram, the message is displayed as:

message(parameter) :returnValue

Instead of as:

returnValue := message(parameter).

Yes, I've double and triple-checked that I am specifying the Return Value in the proper place on the Message Properties dialog.

This isn't correct... is it?


General Board / Re: UML recognised certification
« on: May 17, 2004, 04:54:01 pm »
There are two that I know of:

OMG's OCUP Exam Program,

IBM's Object-Oriented Analysis and Design with UML,

I have the IBM cert, and I think it covers some good material... i.e. it covers more than just knowledge of the UML.  I do not have the cert from the OMG (yet) so I cannot really comment on it.


General Board / Types for MDA Transformations?
« on: June 17, 2005, 10:43:12 am »
What should we use for the “language” or “types” when creating a PIM?  I’m guessing the transformation engine expects OCL types, but I don’t see OCL as a language selection.



General Board / SQL Server Replication for EA repository
« on: March 01, 2005, 03:42:36 am »
Hi all,

We are doing development in both the US and India.  We are using MS SQL Server for our repository and this is located in the US.  We have found EA’s responsiveness when connecting to this repository from India to be flat out terrible.  I’m wondering if it would be possible to use SQL Server replication to replicate the database and improve performance.

Have any of you tried replicating a SQL Server based EA repository?  Did you use Transactional Replication or another type of replication?  Any guidelines or pointers?


General Board / Re: How to get rid of (from package name)
« on: April 20, 2004, 09:24:07 pm »
From the Diagram Properties uncheck "Highlight Foreign Objects".

Best regards,


General Board / Re: Diagram/Project Tree synchronization?
« on: April 21, 2004, 06:49:44 am »
Ah, welcome to the world of modeling vs. drawing.  When you create an element of some kind you are adding that element to your "model," or think of it as your model’s repository.    Once an element is in your model you can display it on multiple diagrams in various contexts.  For example, you might have an Actor displayed on multiple use case diagrams, but it is still the same actor.

Given that, if you deleted an element from one diagram, you probably do not want to delete it from every other diagram!  That is why you need to go through the process of deleting from the project tree itself.

General Board / Re: C# delegate signature declarations
« on: April 21, 2004, 06:36:45 am »
Actually, is it possible to create code templates to handle delegates based on certain stereotypes?  I really have not done that much with the code templates myself so I have no idea.

Pages: [1] 2 3