Book a Demo

Author Topic: Consistency in assigning Classifiers  (Read 11498 times)

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 8626
  • Karma: +259/-129
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Consistency in assigning Classifiers
« on: December 11, 2008, 01:07:43 pm »
[Edit]See following posts... The whole topic is not relevant.  The UI already works as requested...[/Edit]I'm placing this topic here in the UML category prior to submitting a bug report to Sparx.  It's here instead of Bugs and Issues, because there are UML theoretical issues involved first.  I hope to elicit some discussion and provide a considered user view.

When you assign a type to an attribute, you are, essentially, performing the same function as assigning a classifier to an instance.  The only difference is that an attribute can have a type that isn't a classifier.

Indeed, EA actually implements this.  The text of the type in the Type control is persisted to the Type column in t_attribute.  If the Type is an existing Classifier and you've selected it from the drop-down, EA sets the Classifier column to the Object_ID of the selected Type object.

So far so good.  However, assigning a classifier to an instance uses a different control to do this.  In that classifier selection control, you get the ability to filter on namespaces and you also get additional information about the type and location (package) of the classifier you are about to select.  This is to allow you to select the correct one.

My question is:  Why can't we do that for the attributes?

In a serious model, there are often many classifiers with the same name.  In particular, if you are trying to implement OMG multi-level views (CIM, PIM PSM etc) you WILL get name ambiguation.  We do have such a model and it's really starting to hurt...  Sometimes we have as many as 6 entries for the same identifier and no way to distinguish the one we want!

We need the assigning of Classifiers to be consistently presented.  The control engaged when you type [Ctrl+L] should be used for assignment of type to an attribute (and in any other applicable places).  That is, instead of a combo box, there should be a text box and a builder button [...].  Pressing the builder button would engage the Assign Classifier from namespace control.

Thoughts?
Paolo
Consistency, Consistency, Consistency! TM[/s]
« Last Edit: December 12, 2008, 11:36:37 am by PaoloFCantoni »
Inconsistently correct systems DON'T EXIST!
... Therefore, aim for consistency; in the expectation of achieving correctness....
-Semantica-
Helsinki Principle Rules!

Eve

  • EA Administrator
  • EA Guru
  • *****
  • Posts: 8110
  • Karma: +119/-20
    • View Profile
Re: Consistency in assigning Classifiers
« Reply #1 on: December 11, 2008, 05:02:36 pm »
Quote
We need the assigning of Classifiers to be consistently presented.  The control engaged when you type [Ctrl+L] should be used for assignment of type to an attribute (and in any other applicable places).  That is, instead of a combo box, there should be a text box and a builder button [...].  Pressing the builder button would engage the Assign Classifier from namespace control.
Congratulations.  You've almost exactly described how it does work (including for method return types and parameters).  The only difference being that there is a combo box instead of your proposed text box.

So have you just never seen the button that is there, or are you asking for something else?
« Last Edit: December 11, 2008, 05:09:14 pm by simonm »

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 8626
  • Karma: +259/-129
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Re: Consistency in assigning Classifiers
« Reply #2 on: December 12, 2008, 11:33:13 am »
Quote
Congratulations.  You've almost exactly described how it does work (including for method return types and parameters).  The only difference being that there is a combo box instead of your proposed text box.

So have you just never seen the button that is there, or are you asking for something else?
Never "seen" it!  :-[  

I plead "male sight".  That affliction, well known to women, where males can't see what's there...

Thanks Simon!

Well, yesterday wasn't my best day at work and this is another example - a true Homer Simpson moment! :-[

(abashed)
Paolo
Inconsistently correct systems DON'T EXIST!
... Therefore, aim for consistency; in the expectation of achieving correctness....
-Semantica-
Helsinki Principle Rules!