Sparx Systems Forum

Enterprise Architect => General Board => Topic started by: Paolo F Cantoni on August 29, 2016, 05:33:51 pm

Title: Icons ain't icons...
Post by: Paolo F Cantoni on August 29, 2016, 05:33:51 pm
The [X] Use Stereotype Icons checkbox in the Diagram Element Properties is defined as:
For elements that have whole shapes drawn by Enterprise Architect (such as Analysis stereotypes), select this checkbox to draw the alternative shape (if defined).
For elements that have an icon displayed in the top right corner (such as an Artifact element), if 'Show Element Stereotypes' is selected, select this checkbox to display the stereotype icon instead of the stereotype text.

Does this only apply to Artifacts or can it be used to suppress the icon on shapes such as Components?  If so, how?

TIA,
Paolo
Title: Re: Icons ain't icons...
Post by: Paolo F Cantoni on August 30, 2016, 04:50:20 pm
Bump!

Anyone?

Paolo
Title: Re: Icons ain't icons...
Post by: qwerty on August 30, 2016, 07:43:24 pm
Probably Components are not stereotyped classes and so this does not apply here.

Bad design of UML that stereotypes are not stereotypes (e.g. <<enumeration>>).

q.
Title: Re: Icons ain't icons...
Post by: Glassboy on August 31, 2016, 07:47:29 am
Bad design of UML that stereotypes are not stereotypes (e.g. <<enumeration>>).

You're doing that koan thing again (I think I need more coffee).  ::)
Title: Re: Icons ain't icons...
Post by: Eve on August 31, 2016, 08:41:59 am
Bad design of UML that stereotypes are not stereotypes (e.g. <<enumeration>>).

Interestingly, up to UML 2.4.1 in Superstructure Annex B: Keywords, one of the provided reasons for using keywords was "To indicate a Standard Stereotype". It then went on to list the stereotypes from the standard profile in the keyword table. I always generalized this to be "The visual representation of a stereotype being applied is a keyword being displayed."  It disagrees with the UML spec, but it made more sense.

Unfortunately, UML 2.5 Annex C: Keywords no longer includes those points. Instead you'll just find (paraphrased) words between guillements could be either keywords or stereotypes.


Title: Re: Icons ain't icons...
Post by: Eve on August 31, 2016, 08:47:28 am
It doesn't currently apply to components.

Although the UML specification describes that as an icon, so either the help or behavior would appear to be wrong.
Title: Re: Icons ain't icons...
Post by: qwerty on August 31, 2016, 09:13:46 am
Unfortunately, UML 2.5 Annex C: Keywords no longer includes those points. Instead you'll just find (paraphrased) words between guillements could be either keywords or stereotypes.
Don't we all love those ambiguities in standards? ::)

q.
Title: Re: Icons ain't icons...
Post by: Paolo F Cantoni on August 31, 2016, 09:24:02 am
It doesn't currently apply to components.

Although the UML specification describes that as an icon, so either the help or behavior would appear to be wrong.
Could it?  Should it?

Should I pop in a feature request?

We currently have to create an overriding rectangle to hide the "icon" and then place our own over the top for certain - but as usual with EA not all - base types for our MDG.  If the user selects Gradient fill, the appearance of the shape is "scungy" (1).  If some "icons" can be hidden, why not all?

Paolo

(1)Technical term...  ;D