11
« on: November 17, 2023, 10:09:07 am »
Hi forum, on behalf of my organization, I am posting a feature proposal to see if there are interest in such a functionality and or other ideas you may see valuable in this context. The feature is about enhancing or replacing the not so user friendly “Virtual connector” trick.
We have encountered a challenge in the current version of Enterprise Architect (as of November 2023): there is no direct functionality that allows for the multiple visual representation of a class (or element) within the same diagram (except for the virtual connector trick). Such functionality would be crucial for simplifying diagrams where central classes are connected to many others, creating a visually complex network of links that can reduce clarity and hinder understanding.
I propose the addition of a feature that would enable users to create multiple visual representations of a class within the same diagram. These multiple representations would act as alternate views of the same element and would not be treated as distinct elements within the model. The primary aim here is to enhance diagram readability and the traceability of class/element relationships without altering the integrity of the underlying data model (or any other types of models or diagrams).
Ideally, each representation of the same element on a diagram could also have it own set of visual properties (color, font, size, visible compartments, etc.).
It also means that we can select each occurrence individually on the diagram. Whereas this is not possible with the current implementation of the not so friendly virtual connector trick. And may be a set of ctrl keys combined with the element selection click could select all representations of the same element on the diagram.
Adding a new relationship to one of those visual representations element on the diagram will add the connector to the real elements but the new link will only appear between the two element representations at both ends of the new relationship (source and target elements).
We have now to think about what will happen on other diagrams where the same elements are present.
There are 2 cases:
1. There are no multiple representations of the corresponding elements
As usual, the link is visually added on the diagram between the elements accordingly to the connector’s properties of the diagram: Show Relationships, Freeze Visible
2. There are multiple representations of the corresponding elements on the diagram
The link is added between only one of the multiple representations producing the shortest path or the path that cross the least number of relationships.
With the additional possibility to manually reassign the link to another representation element if desired.
In order to maintain clarity with these multiple representations, we suggest an intuitive labeling system: if a class, let's say "Class A", is represented 3 times on the same diagram, the occurrences could be labeled as "Class A (1/3)", "Class A (2/3)", and "Class A (3/3)" respectively. This would indicate that they are visual copies and would show the total number of instances displayed in the diagram.
This approach ensures that users can quickly discern that these are not different and unique classes but different representations of the same class, thereby maintaining model integrity while improving understandability.
What do you think of this proposition ?
Should I send a feature request to Sparx ?
Show Posts
