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.


Topics - sargasso

Pages: 1 ... 6 7 [8] 9 10 11
106
General Board / Some strangeness in the header
« on: April 10, 2005, 10:37:58 pm »
Two C# classes, both with exactly the same attributes.  One generates code with the line
    using Arrays;
in the "header".

The other doesn't.  Any clues????

YIPB
bruce


107
General Board / Hiding inherited constructors
« on: March 23, 2005, 04:10:30 pm »
Hi,

Looking at C#.

I want to show the inherited methods for a class element in a diagram, with the exception of the parent heirarchy's constructors. All constructors across the whole model are streotyped as "ctor".

If I use the context menu Element Features|Specify Feature Visibility and enter "ctor" EA hides all the constructors, including the current class constructors (I want to be able to see these!).

is there a way of hiding by stereotype on inherited features only? :-/

bruce

108
General Board / Reeeeally small diagrams (again)
« on: March 08, 2005, 05:40:04 pm »
I've had to regenrate my base project template and now whenever I produce an rtf report - any report - the diagrams are really really tiny.  I need to change the size up by a minumum of 1000% before they are viewable.

I had this problem a long time ago and found the solution.  now I've forgottent it again.  

Does anyone know the answer?

bruce

109
General Board / Multiple lifecyles on a timeline
« on: March 01, 2005, 07:15:38 pm »
OK I think I'm getting the hang of these dang "alt" and "ref" fragment thingies.  

But, ... 2 questions.

I want to  show the two outcomes of two alt flows both of which have terminating effects on the lifecycle of one of the objects.

In the first case, the object sends a (asynch) message to an object not of concern to this model and then exits.  So, I can either
1) treat is a "lost" message and set it between the object and an Endpoint - fine. ..or
2) add a ref fragment and a message gate and point it there - equally fine.
Now here's the problem.  Once that message has been sent, this object can exit.  IOW the lifecycle (not just the activation level or of the source object ends.  I dont see how to make this happen?  I have even tried using a self message but I still dont get the little X to appear.  The only way to get it is to create an incoming message?

Now the second case.  In this alt fragment something happens that just makes the object exit - let's say user clicks "Cancel" button - without doing anything else.  Now, this one works - sort of.  I add the message from the user to the object and set the lifecycle to delete.  OK now I get the little X under the lifeline end.  But, this now overrides the X in the above fragment. ???

Is dis a bug?

bruce

110
General Board / Is it me, or is there something wrong with UML
« on: February 15, 2005, 07:29:42 pm »
OK, I've got a web app here that involves a whole bunch of components:
The client browser, 2 web servers, 2 app servers,  a three tiered back end system, 2 external payment gateways (one, CCD validation, is hidden from the web user and one, deferred payment service from the PO, involves direct user interation) and the customers e-mail clientd (we can ignore the SMTP server at present).

What I am trying to do is model all the messages between the components that make up a transaction.  Simple heh?

The EA elements I am trying to use all exist as components in the Component and Deployment models - as components - and they all have vast depth of use in the rest of the repository.

Now ....

Are components classifiers or instances??????

(Actually I'm not too concerned about that.)


Now,  lets say I want to start modelling the tx flow at "User accepts payment terms by clicking on the submit button on the Buy Now page on the browser".  
This would generate a URI request to the payments web server relating to a specific script in the payments app server which would submit a credit validation message to the CCD gateway which, if credit is accepted, would send a n acceptance message back to the payment app which would then raise an order by sending a message to the back end comms tier which would create an order in the sales / warehousing system  which if succesful would return an order number back through the tiers to the app server which would then create a payment record against the order through the same tiered ..... blah blah blah  ....  which would finally send a confirmation email to the customer.   :-X

Now all these components have assembly links that currently describe the protocols and mechanics of sending messages between the two.

I want to add the messages ...  :'(

Seems I got to create associations between the components as that is the only way to add messages.

Is this correct?  i.e. doesn't UML like communications messages on assembly links

or

Is EA not doing it?



yours in misery

bruce

111
General Board / Computed properties - a question and a bug
« on: February 08, 2005, 02:13:02 pm »
Ver 4.50.746 (beta)

Question:  Some short time ago there was a discussion of whether properties should be treated differently to methods in class diagrams.  I have tried various searches but cant find it.  Does anyone remember what the topic was?

Bug(s): Let P be a class property whose value is computed - i.e. there is no associated private attribute.  Reason dictates that it is a readonly property.  When I RevEng it from C# EA creates the correct representation:

<<property>> int P with tag readonly=true and no associated attribute.

Now here's the twist - we have now decided to expose a setter for the property. (Why is complex - let's just say we want to persist a value passed in)

I have tried both setting the readonly tag to false and deleting the tag.  On forward engineering - back into the existing source - the setter method is not added.

Can anyone see why this would not be a bug?


tia
bruce

112
General Board / Find all "diagram" elements, or...
« on: January 31, 2005, 02:47:55 pm »
OK, having created a bit of a mess by splitting a use case into two new use cases (each of which has an associated activity diagram) I need to find all the places in this model where I have inserted the activity diagram from the original use case as a "hyperlink" so I can change the relevant ones to the new diagram.
(OK - I admit it - I have actually split 15 use cases.... :( )

Does anyone know how I can locate all the diagrams that have these hyperlinks in them.

tia
Bruce

113
General Board / Back button for the project browser
« on: January 20, 2005, 02:03:19 pm »
Good morning campers!

Is there any way I can set up a "back" button that works in in the project browser.

Strange question, I know, but I am currently working through the requirement -> usecase -> test case maze and I would like to
1) pick the next requirement (in the requirements package) in the docked project browser.
2) pick the appropriate use case(s) from the realtionships browser (docked)
3) use the context menu to locate the related object (ie the use case) and then....
4) look through the use case tests to see if we have adequately covered the reauirement.
5) (not yelling just highlighting) MAKE THE PROJECT BROWSER GO BACK TO THE REQUIREMENT I WAS JUST LOOKING AT.

In order to do this I have the project browser, the element properties window, the relationships window and the element Notes window all docked and therefore about 1mm of screen width for the diagram.

114
General Board / Search in tests
« on: January 11, 2005, 03:51:53 pm »
Is there any way to search for a string in element tests?  :-/

As in , I know which usecase the test I'm looking for is in but the string - which could be in the name or the description -  is  a particular fare type.  The use case "book travel" has over 200 test cases  ....

tia
Bruce

115
General Board / Lost in space
« on: December 19, 2004, 03:47:46 pm »
On the project browser context menu is a wonderful item "Create Link..." which lets me create a link between 2 elements without having to put the two elements in a diagram.

My problem is - where is this damn item in the keyboard customisation lists?

Alternatively, is there a way I can add a link from/to an element in a diagram without having the other element in the diagram.  What I am trying to do is to create the requirement realisation links to the use cases.  The requirements were sucked in to the model as external requirements, they all live happily in there own package and I dont want them in the UC diagrams - but I do want to create the links.


Frustrated....

116
General Board / Change an Activity Initial to a Synch
« on: December 14, 2004, 01:28:29 pm »
Activity diagrams.

Now that the entire model is coming together, it is now apparent that some of the start nodes on the flows shouldhave been modelled as synch elements - a pedantic need but nonetheless...

But how can I change them?  The Change element type function treats them all as "StateNode"  and I cant see anything that differentiates them in the properties.

Any ideas?

tia
Bruce

p.s. due to SOE we dont have access to Access  :(

117
General Board / Configuration/Maintenance/Problem Types
« on: December 07, 2004, 03:58:19 pm »
Where are the problem types defined in the Config menu used?

From the Help "For the maintenance and change control screens, you can set the base Problem Types that are handled. "

but I cant find it...   :(

I just spent 2 days discussing and setting up 39 sev levels for this project ....

Bruce

118
General Board / New tagged value handling
« on: November 30, 2004, 05:38:07 pm »
How can I access tagged values for elements opened from
1. the search window.
2. the relationship matrix
3. as external requirements from and opened element.

In fact unless the  element is in the current diagram I cant get to its tagged values.  V. nasty for elements not in diagrams.

>:(
Bruce

119
General Board / Objectnodes on activities
« on: December 02, 2004, 06:39:39 pm »
Is there a way to turn off the default behaviour of automatically adding objectnodes to object flows between activities?

(searched help and here - no joy)

Bruce

120
General Board / Multidirectional control flows
« on: November 02, 2004, 07:43:43 pm »
I am stuck on this one.

I have an activity diagram for a web app.  At two points in the transaction flow the user can invoke a popup window that shows some stuff and allows them to select an option.
(In fact what it is, is a train fare class select page that alows them to pick a class and view the price and options for that class. The mainstream pages have a dropdown to select an initial class, they can then use these popups to peruse and select different classes and options and see the price effect.)
The popup page returns the set of options last selected when they click on the "Close" button.
Now this is my problem.  There are two inflows to this activity and when they click close the flow must unambigously go back to the activity they were in when the invoked it.
Control flows (by definition) are unidirectional.  So how do I show the returns must go to "where they came from"?

rgrds
Bruce


:-[ Oh, never mind I just realised thats what guard conditions are for  :-[

Pages: 1 ... 6 7 [8] 9 10 11