Archimate 3: Cannot Create an assignment between a device and a system software

[Uli Schulze-Eyssing]

... but should be possible.
Do I miss somthing?
Many Thank, Uli

[Sunshine]

Nah your not missing anything. Its just broken.
Glad I created my own MDG for ArchiMate 

[matthew.james]

What version of Sparx are you running?

There are definitely many valid Archimate relationships that are considered invalid by Sparx.  But (at least on 14.0 Build 1420) I can create an Archimate_Assignment relationship between a Archimate_Device element and a Archimate_SystemSoftware element (in either direction).

In case you are not aware, you can disable the validation by turning off "Strict Connector Syntax" - although keep in mind that this an all or nothing switch.
http://www.sparxsystems.com/enterprise_architect_user_guide/14.0/user_interface/links.html

[Glassboy]

I can create an assignment both ways.

[Sunshine]

Archimate 3 MDG seems to be broken in build 1427.

In case you are not aware, you can disable the validation by turning off "Strict Connector Syntax" - although keep in mind that this an all or nothing switch.
http://www.sparxsystems.com/enterprise_architect_user_guide/14.0/user_interface/links.html

Thanks I didn't know that.
One of the benefits of participating in this forum is learning new stuff aye?

[matthew.james]

Archimate 3 MDG seems to be broken in build 1427.

Hmm - thanks for the warning.  I'll be holding off from upgrading then

[Uli Schulze-Eyssing]

My Version is 14.1.1427

[Graham_Moir]

My Version is 14.1.1427

Have you logged a formal bug report referencing this thread?

[Sunshine]

My Version is 14.1.1427

You could download an earlier version and use that instead or alternatively use Archimate 2

[Eve]

Archimate 3 MDG seems to be broken in build 1427.

In my opinion, it's the ArchiMate 3 specification that's broken. It's self-contradictory and ambiguous.

EA has implemented the relationships as defined the the metamodel described by the specification. We are aware that it's different than the appendix, which attempts to provide an interpreted view of that metamodel.

[matthew.james]

Simon

That's interesting, and useful to know.  I had presumed that the appendix (by which I assume you mean "Appendix B  Relationship Tables"?), being published with the standard, is part of the standard and was added to clear up ambiguity by making all allowable relationships explicit. However I am far from being an expert in Archimate and I haven't read through and cross-validated the appendix against each elements' specific definitions; beyond noticing that many (all?) of them note that the metamodel diagram doesn't show all permitted relationships.

Can I take it then that, when trying to implement the relationship rules, the Sparx team have found relationships listed in the appendix that are explicitly precluded within the detailed specification? Do you know of any off the top of your head to give me a pointer into what to look out for? Do you know if these issues have been reported to Open Group?

Thanks

[Eve]

Matthew, it's a little more complicated than that.

In addition to the metamodel, there is a textual description of how derived relationships work. This is where the "ambiguous" comes in to my comments. However, even taking that into account there are some connectors allowed by the table that I can't see how they fit into the metamodel. EA currently doesn't have those derived relationships added. Most issues reported come down to that.

There's also a comment somewhere about something not happening between layers. Going from memory here, so I don't have the specifics. But the effect of that comment is excluding relationships in the table that the metamodel would otherwise allow.

EA 14.1 uses a metamodel approach to build quicklinkers for UML, UAF, ArchiMate and user technologies if they choose. It works well when the metamodel is well defined. Unfortunately, it doesn't necessarily meet user expectations when the specification writers have ignored the metamodel and defined a table showing what the available relationships are.

[Sunshine]

Yes there were problems with the relationships in the ArchiMate V3.0 spec. Probably because they tried to define the metamodel without a tool - sigh  . I suspect the authors of the spec ended up having mental fatigue defining the relationship tables and ended up publishing V3.0 without the necessary vigour they should have done. But hey they are human and everyone has bad days.

However after publishing I understand that there were some articles pointing out those errors by Gerben Wierda and a few others lead to V3.01 of the ArchiMate Spec which addressed most of those issues. Not saying they got it perfect mind.

So I guess the question is does the ArchiMate MDG take into account those changes in V3.01. I suspect perhaps not given this thread.
The simple solution would be to revise the ArchMate MDG to comply with V3.01 then things might work a little better.

[Geert Bellekens]

Archimate 3 MDG seems to be broken in build 1427.

In my opinion, it's the ArchiMate 3 specification that's broken. It's self-contradictory and ambiguous.

EA has implemented the relationships as defined the the metamodel described by the specification. We are aware that it's different than the appendix, which attempts to provide an interpreted view of that metamodel.

Simon,

I'm not saying you are wrong, I'm not a fan of the Archimate specification either, but in this particular case that can't be an excuse.
Figure 77 of the ArchiMate metamodel clearly shows the relation between device and system software.

Geert

[Eve]

Simon,

I'm not saying you are wrong, I'm not a fan of the Archimate specification either, but in this particular case that can't be an excuse.
Figure 77 of the ArchiMate metamodel clearly shows the relation between device and system software.

Geert

You're right. I remember a report about that particular issue and adding it in to our metamodel now. If you don't want to wait for a release, add the text in bold to your ArchiMate3 technology file. (The rest is for context)

          <Stereotype name="ArchiMate_Device" metatype="Device" notes="" bgcolor="12648384" fontcolor="-1" bordercolor="-1" borderwidth="1" hideicon="0" generalizes="Node" baseStereotypes="Node" strictness="profile">
            <stereotypedrelationships>
              <stereotypedrelationship stereotype="ArchiMate3::ArchiMate_Assignment" constraint="ArchiMate3::ArchiMate_SystemSoftware;ArchiMate3::Event;ArchiMate3::Internal Behavior"/>
            </stereotypedrelationships>
            <Icon type="bitmap" xmlns:dt="urn:schemas-microsoft-com:datatypes" dt:dt="bin.base64">

System software was previously inheriting the stereotyped relationship to Event and Internal Behavior. Because the stereotyped relationship overrides any existing relationships of that type we need to add them too.