Messages

1

General Board / Limit package contents displayed

« on: November 11, 2015, 10:52:05 pm »

Hi,

I'm creating a package diagram and am showing the contents of the packages, but I want to limit the contents that are displayed.

I keep all my Issues in Issues packages and have stereotyped them as 'IssuesPackage'. I then right-click on the package I'm displaying, select 'Feature & Compartment Visibility' and add IssuesPackage to the list of stereotyped features to hide. However, the Issues package is still visible.

I suppose I have two questions:
1. Are package contents not package features and therefore this stereotype hiding doesn't apply to contents?

2. If so, is there another way to constrain the displayed package contents?

Many thanks in advance for any help.

Michael

2

General Board / Re: Component vs classes

« on: July 15, 2015, 01:57:06 am »

Found it now, thanks.

I generally assume (maybe incorrectly) that if EA won't let me do something, then it's not supported in the UML spec.

3

General Board / Re: Component vs classes

« on: July 15, 2015, 01:17:54 am »

Hi,

Thanks for your replies. I have seen the example model - specifically the component diagram in the UML section. Confusingly though, this does show aggregation associations - one even looks like composition!? But I can't seem to generate these type of associations myself between the same objects.

On the book front, unfortunately the ones I've read, including online resources, are as unhelpful as the last comment. They mention that a component generally has a larger responsibility or scope than a class (which is why I'm using them), but don't describe why I can't show composition and/or aggregation like I can with classes, specifically as I need to show multiplicity.

4

General Board / Component vs classes

« on: July 10, 2015, 08:54:49 pm »

Hi,

I'm developing a system model and as a non-developer I tend to be interested in applications, hardware or just logical "things" rather than a class meaning an actual OO class. So I've set out doing most of my modelling using components.

However, every now and then I encounter something which makes me think I should maybe use classes instead.

Specifically, EA doesn't allow me to use composition and aggregation associations between components. I can show a single component inside another component (assuming this is the closest thing to composition) but I want to be able to show multiplicity. e.g. server A comprises several instances of application A each of which talks to single application B. I realise I might have to separate my structure from my messaging (I'm actually keeping messaging more abstract and just using associations labelled "sends messages to"), but is a component diagram the best way of showing these things? Or, if I want composition multiplicities, should I be using classes?

Sorry - a bit rambly and maybe unclear! Please let me know.

Appreciate any help with this.

Thanks,

Michael

5

General Board / Re: Showing child components on a sequence diagram

« on: June 25, 2015, 11:08:40 pm »

Thanks for your reply - at least I can stop looking for that functionality!

6

General Board / Showing child components on a sequence diagram

« on: June 24, 2015, 02:28:21 am »

Hi,

I'm drawing a sequence diagram for component interactions and I'm trying to show the internal messaging for one of the components. I know you can use ports and they sit nicely underneath the parent component, but is it possible to have child components shown in the same way? i.e. attached to and underneath the parent?

I think you can do this in SysML, where 'composed of' blocks are shown underneath the composing block, but I can't find anything analogous in UML.

I'm using components for all my objects which can't use composition/aggregation for a start - perhaps I'm mixing my semantics up.

Any though on how to show "child" components of a component doing intra-component messaging on a sequence diagram would be most welcome!

Many thanks in advance,

Michael

7

General Board / Use Case Details documentation template

« on: April 01, 2015, 07:57:00 pm »

Hi,

I'm trying to use the stock 'Use Case Details' template to generate some documentation on a bunch of use cases I have. The trouble is that when the document is generated, the scenario fields are not populated and appear in the generated document still in template form:

{ElemScenario.Notes}

structured scenarios >
{Scenario_Structured.Step}. {Scenario_Structured.Action}
Uses:
{Scenario_Structured.Uses}

exception >
{Exception.Type}:  {Exception.Step}.    {Exception.Name}
< exception
< structured scenarios

I'm using EA 11.

Has anyone else come across this, or might be able to point out what I might be doing wrong?

Thanks very much for any help.

Michael

8

General Board / Specification Manager - filter by keywords

« on: March 06, 2015, 04:14:43 am »

Hi,

I'm trying to find a nice way of filtering requirements. I thought I'd use keywords so that I can use them in conjunction with diagram filters to highlight objects of interest.

However, I'd also like to filter by keyword in the Specification Manager, but this doesn't seem to be a column I can add. Is there any way around this, perhaps by configuring the list of available columns?

Thanks in advance for any help.

Michael

9

General Board / Re: Relationship Matrix and traceability

« on: February 27, 2015, 03:01:50 am »

Thanks very much for your reply, I'll figure out another way of doing this.

10

General Board / Relationship Matrix and traceability

« on: February 26, 2015, 11:20:57 pm »

Hi,

I'm using the Relationship Matrix to map requirements to components, test cases etc.

I've set up two tiers of requirements - we have a whole bunch of customer requirements from different customers and then a rationalized set of product requirements.

I've currently mapped product requirements to one or more customer requirements. The product requirements are also then mapped to test cases with a trace association.

The trouble I'm having is using the Relationship Matrix to see the indirect relationships between a customer requirement and a test case:

customer reqt <-> product reqt <-> test case

I presume this is down to the Relationship Matrix only supporting one-hop relationships? Is there any way of achieving this if I were to change the type of relationships? For example, if I made the customer reqt a specialization of the product reqt? (tried that, didn't work   )

Maybe it's not possible, just wanted to see if it might be. Any help would be very much appreciated.

Thanks,

Michael