Shapescript vs Local diagram rendering

[Paolo F Cantoni]

My company tries to do things with color as well, no matter how much I tell them not to.

Color of model elements in a diagram is nice, but has no semantic value. It's not "part of the model".  You should always use stereotypes or tagged values when giving different elements different "meanings".   If you want to give color to stereotypes or tagged values for a visual effect, than that's fine, but never use color alone.

Your shape script is drawing based on stereotype or property (ie, tagged value), just as it should.  Use a tagged value to color your elements.

We're trying NOT to, Sparx EA forces us to.  We are using colour as a PROXY for the diagram specific properties that conceptually should be there.

Notwithstanding that diagrams are NOTIONALLY merely views of the underlying model, as Sparx has discovered (at least, so far, only for vertices) there may be properties that are conditional on specific views - sort of like conditional tagged values.

In the absence of such functionality, we're FORCED to use actual rendering properties (such as colour) as proxies for the conceptualisation.

Paolo

[Geert Bellekens]

My company tries to do things with color as well, no matter how much I tell them not to.

Color of model elements in a diagram is nice, but has no semantic value. It's not "part of the model".  You should always use stereotypes or tagged values when giving different elements different "meanings".   If you want to give color to stereotypes or tagged values for a visual effect, than that's fine, but never use color alone.

Your shape script is drawing based on stereotype or property (ie, tagged value), just as it should.  Use a tagged value to color your elements.

I couldn't agree more.

Geert