Author Topic: Suppress line segment behaviour anomalous (bizarre)  (Read 443 times)

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 6801
  • Karma: +142/-103
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Suppress line segment behaviour anomalous (bizarre)
« on: April 15, 2019, 05:48:47 pm »
When you suppress a line segment, the name of the arc (if there is one) is used as the label for the end-points of non-suppressed segments.  If the name is of any substantial length, this will "b*gg*r up" your diagram almost as much as the non-suppression of the line segment in the first place.

While it is appropriate for the name of the arc to be used as the default value of the end-point identifiers, it is not valid to make the name of the arc and the identifiers for the end-points one and the same!  Our users have been blithely stuffing up other peoples' diagrams because it wasn't understood (by any of us - myself included) that setting the end-points wiped the name of the arc.  Also, the value on one diagram was duplicated on another!  Confusion abounds as a diagram can end up with 5x"A"!  Then the users end up with "duelling End-points".

<rhetorical question>Can anyone provide a rationale for this totally less-than-optimal behaviour?</rhetorical question>

The suppressed end-point identifiers have to be diagram specific, not global.

It's a bug, ...

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

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 6801
  • Karma: +142/-103
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Re: Suppress line segment behaviour anomalous (bizarre)
« Reply #1 on: April 24, 2019, 05:05:22 pm »
I got a response from Sparx.  No rationale, just "it is what it is and it's not a bug".

I replied that the functionality that is being implemented pre-existed Sparx EA, pre-existed computer graphics (indeed pre-existed my learning to draw engineering diagrams at Uni - with REAL pens and pencils!) and that the current implementation violates many of the requirements that stem from the real world functionality.   Consequently, it is a user experience defect.

I'm still open to any rationale for the existing behaviour.

Paolo
« Last Edit: April 24, 2019, 05:09:19 pm by Paolo F Cantoni »
Inconsistently correct systems DON'T EXIST!
... Therefore, aim for consistency; in the expectation of achieving correctness....
-Semantica-
Helsinki Principle Rules!

MaXyM

  • EA User
  • **
  • Posts: 120
  • Karma: +7/-0
    • View Profile
Re: Suppress line segment behaviour anomalous (bizarre)
« Reply #2 on: April 24, 2019, 07:58:46 pm »
I can agree with you. Doesn't make sense.
I can agree that it's not a bug (as programming mistake). Just bad or incomplete design. I found it's trendy recently, to justify unexpected software behavior to call "design decision". It's hard to argue against, even if common sense of the feature doesn't fit expectations.

In overall this feature seems to be rather unfinished. Using it a few times unveiled several glitches. I didn't even bother to report it.
Not to mention that 'hide segment' feature is available in context menus for connections which are not applicable for.

With regards

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 6801
  • Karma: +142/-103
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Re: Suppress line segment behaviour anomalous (bizarre)
« Reply #3 on: April 25, 2019, 10:54:28 am »
I can agree with you. Doesn't make sense.
I can agree that it's not a bug (as programming mistake). Just bad or incomplete design. I found it's trendy recently, to justify unexpected software behavior to call "design decision". It's hard to argue against, even if common sense of the feature doesn't fit expectations.

In overall this feature seems to be rather unfinished. Using it a few times unveiled several glitches. I didn't even bother to report it.
Not to mention that 'hide segment' feature is available in context menus for connections which are not applicable for.

With regards
It's a defect.  You can't tell if it was a correctly coded design defect or a badly coded good design.  In my view, reserving the word bug for solely a coding defect is too limiting.  Besides the phrase "it's a defect, ..." doesn't sound as emphatic...   ;D

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

Glassboy

  • EA Practitioner
  • ***
  • Posts: 1301
  • Karma: +103/-75
    • View Profile
Re: Suppress line segment behaviour anomalous (bizarre)
« Reply #4 on: April 29, 2019, 07:58:17 am »
I can agree that it's not a bug (as programming mistake). Just bad or incomplete design. I found it's trendy recently, to justify unexpected software behavior to call "design decision". It's hard to argue against, even if common sense of the feature doesn't fit expectations.

Well No.  It doesn't fit your current expectations.  At some stage it probably fit someone else's now past expectations.  That's why all design decisions need to be fully documented so their context and rationale are understood.  So you know when things change and a new design decision is needed.  Or you tell someone to bugger off because they represent an uneconomic corner case.

MaXyM

  • EA User
  • **
  • Posts: 120
  • Karma: +7/-0
    • View Profile
Re: Suppress line segment behaviour anomalous (bizarre)
« Reply #5 on: April 29, 2019, 07:37:12 pm »
Quote
At some stage it probably fit someone else's now past expectations.
Sure, PMO's expectation.

Quote
That's why all design decisions need to be fully documented

And that's why are not being documented: to not give a proof that low quality was a design decision.
In fact common its practice in big companies.

BTW I wouldn't call a 'corner case' to distributing name of a connector across a whole model.

Glassboy

  • EA Practitioner
  • ***
  • Posts: 1301
  • Karma: +103/-75
    • View Profile
Re: Suppress line segment behaviour anomalous (bizarre)
« Reply #6 on: April 30, 2019, 07:20:12 am »
BTW I wouldn't call a 'corner case' to distributing name of a connector across a whole model.

Anything not currently bugging me about the product is a corner case :-)