SysML: How to setup EA?

[Matthias Weber]

I want to model systems with SysML1.1, so i tried to setup EA for using SysML 1.1 only. It seems not to be possible, since the whole IDE is "poluted" with things which aren't part of SysML but are part of UML (e.g. certain stereotypes, attribtues).

Is there any chance to configure the IDE so that it is pure SysML?

Bye
Matthias

[Geert Bellekens]

Matthias,

Have you tried to uncheck all other MDG technologies except SysML?

(go to "Settings/MDG Technologies")

I think that should help

Geert

[Matthias Weber]

yes, i already removed any unnessecary things.

[Geert Bellekens]

Then I don't think there's much else you can do.

Geert

[Paul Lotz]

OK, a comment first.  SysML is not independent of UML in any case.  See Figure 1 here: http://www.omgsysml.org/.

What you can do is select the SysML toolbox for each diagram.

Perhaps you want more?  I want to be able to switch between SysML and UML since I want to use them together.

[fwoolz]

I agree with Paul. There is no reason why a system model can't be composed of SysML and "vanilla" UML; in fact, I've done just that in several projects (one of them involving an IEEE standard).

I do think, however, that EA needs to do a better job of distinguishing between stereotypes based on namespace. At present (or at least the last time I checked!) there are numerous name collisions that can lead to application of the wrong stereotype with the right name.

[Paolo F Cantoni]

I agree with Paul. There is no reason why a system model can't be composed of SysML and "vanilla" UML; in fact, I've done just that in several projects (one of them involving an IEEE standard).

I do think, however, that EA needs to do a better job of distinguishing between stereotypes based on namespace. At present (or at least the last time I checked!) there are numerous name collisions that can lead to application of the wrong stereotype with the right name.

Hi Fred,

I had a recent response (from Sparx, on a related matter) that indicated that EA now retains a "memory" of where the stereotype, when applied, came from.  I asked some questions regarding EA's behaviour under certain scenarios, but haven't heard back yet.

I might send them a reminder...

HTH,
Paolo

[Matthias Weber]

I understand SysML as a fully defined language, which can be used to model any technical system completely. Therefore it is IMHO not important, that SysML intersects UML. If one wants to use SysML and UML together, EA should support this optionally.

The reasons why I prefer a clean tool for SysML are:
1. the superset of all systems, which can be modelled with SysML, also contains those, which do not consist of any piece of software.

2. system engineering in large companies is done mostly by classical engineers not by software developers. Those people usually have a rather limited ability for abstraction (hurdle 1) and no knowledge about software development processes or UML (hurdle 2). The approach of MBSE is a huge break with common engineering methods. To bring those people close to MBSE among other things a tool is needed with a minimum of complexity, which can be learned/used easily and hence reduces the risks of maloperation and integration of errors into system models. Any "unneccessary" stuff flattens the learning curve and therefore bears the risk of refusing the tool in the first place and consequently the method by the staff.

[Paolo F Cantoni]

[size=18]...[/size]
The reasons why I prefer a clean tool for SysML are:
1. the superset of all systems, which can be modelled with SysML, also contains those, which do not consist of any piece of software.
[size=18]...[/size]

The set of all systems which can be modelled with SysML, surely Mathias?  

But you make some important points...

Paolo