Connectors not visible after ActiveX import

[sitati]

Hi all,

I'm having an issue that hopefully someone can help me with. I've used a VBA macro to import user stories and connectors from an Excel worksheet. After the import, the connectors are visible in the relationships window, but do not show up in the relationship matrix and traceability windows.

I originally encountered this on a mysql hosted project, and assumed it was a security/locking issue, but I can reproduce the same behaviour on a local project so I am stuck at the moment.

I am using EA 8.0.863, and a snippet of the connector-creating code is below

Set conn = fromStory.connectors.AddNew(connectorName, connectorType)
conn.ClientID = toElement.ElementID
conn.Type = connectorType
conn.name = connectorName
conn.SupplierID = fromStory.ElementID
If direction = "" Then
'default: source to destination
direction = "Source -> Destination"
End If
conn.direction = direction
conn.Update
fromStory.Update
toElement.Update
'refresh the contents of the package so that the changes are seen EA
eaConn.refreshModelView eaConn.getSelectedPackage

Thanks in advance, and appologies if this has already been addressed, I did try to search the forum before posting  

[Geert Bellekens]

Things to check:
- Do your connectors show up after you close and reopen the whole model
- Do they show up when you manually create a connector of that type?

Geert

[sitati]

Hi Geert, thanks for your reply.

1. No, the connections do not appear after reopening.
2. Yes, links made inside EA appear in the matrix and traceability tools as normal.

My first assumption was that it's something wrong with my code, but isn't it strange that the links do appear in one place (Relationships window) but not in others?

[sitati]

Fixed it! It was indeed my code, and quite a silly bug really: I was adding the connectors as 'Realization' links instead of 'Realisation' links. I guess I'm too used to assuming US spelling in programming languages, and EA actually spells the word correctly.

EA will let you add links that are not named in t_connectortypes and show that these connections exist, but they won't show up in traceability (because EA doesn't know how to interpret them) and won't show up in the matrix (because these foreign links do not fall under any of the available connection types).

Thanks for your help.

[Paolo F Cantoni]

Fixed it! It was indeed my code, and quite a silly bug really: I was adding the connectors as 'Realization' links instead of 'Realisation' links. I guess I'm too used to assuming US spelling in programming languages, and EA actually spells the word correctly.

No it doesn't...  And it varies with the version - see: http://www.sparxsystems.com/cgi-bin/yabb/YaBB.cgi?num=1253503550/0

What version are you running?  I thought the dichotomous inconsistency had been fixed in the latest versions...

The Sparxians can confirm...

Paolo

[fwoolz]

Paolo,

Nopenopenopenope.... not fixed. 8.0.863 still has "Realisation" in t_connectortypes. The latest UML standard, however, sez:

7.3.45 Realization (from Dependencies)

The correct spelling is whatever the damned spec says it is! And EA is, therefore, non-compliant in this area.

This may seem like a minor point, but it has generated its share of posts and, obviously, continues to trip people up.

[KP]

The correct spelling is whatever the darned spec says it is!

It certainly isn't spelt "Realization" in the Japanese edition of EA. Are the Japanese doing it wrong then?

The internal representation of our data in our databases is none of the OMG's business. EA has had a "Realisation" connector since long before I joined Sparx, and while we provide an internationalisation layer between the database and the UI that translates "Realisation" to "Realization" or whatever the Japanese equivalent is wherever it's seen, if you access the database directly you will unfortunately bypass the internationalisation layer and see the untranslated data. I'm afraid there's not a lot we can do about that.

[KP]

Of course, it gets a lot more complicated when you add in ArchiMate's "Realisation" relationship which mustn't be translated to "Realization". That introduced the bugs which Paolo was referring to and which were fixed for EA 7.5.849 last October.

[KP]

PPS...

"I don't give a darn for a man that can only spell a word one way." -- Mark Twain



EDIT: And YaBB has just translated Twain's original d-a-m-n to darn to avoid offence. It's almost as if it knew!

[fwoolz]

Neil,

Point taken regarding Japanese... but we speak the same flippin' language, mate!

Whilst we're discussing internationalization and internal consistency, why do we find this (from t_propertytypes in one of my projects)?

Property	Description	Notes
associatedColor	Color associated with an enumeration value	Type=Color;

Would EA accept "Colour" as well?

Cheers,
Fred

[Geert Bellekens]

Oh nooes, not the Realisation vs Realization issue again

Geert

[Paolo F Cantoni]

Point taken regarding Japanese... but we speak the same flippin' language, mate!
[size=18]...[/size]

Not according to Oscar Wilde - who, when speaking of the US and the UK, opined:  "Two countries separated by a common language".

In my original post I flagged above, I was only concerned about the self-inconsistency.

At the time, it wasn't revealed that the Archimate "Realisation" relationship was the cause of the bug.  Interesting!

Anybody know why one part of OMG (Archimate) decided to use Realisation and another part (UML) use Realization - on the assumption that they are both talking about the same concept...

Paolo

[Paolo F Cantoni]

[size=18]...[/size]Anybody know why one part of OMG (Archimate) decided to use Realisation and another part (UML) use Realization - on the assumption that they are both talking about the same concept...

Paolo

Further research shows that the Archimate 1.0 Technical Specification which is from Open Group - NOT OMG (my bad).  Uses Realization and Generalization (as per UML).  The publications I was referencing earlier were using a different language.

Hopefully this clarifies (and negates Neil's point?).
Paolo