Skip to content

Feed aggregator

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Thu, 04/25/2013 - 05:10
Responding to Chriss...

This is the verb form. The noun form refers to the high level,
conceptual view of the actual solution to some problem.

This sounds like a Structured Programming view from the '70s, where
analysis was really part of defining requirements (i.e., you used
analysis to define what the actual problem to be solved was). In an OO

Categories: Communities

Re: understanding class/classifier,modeling/design

UML Forum Google Group - Thu, 04/25/2013 - 05:10
Analysis models are partial descriptions of a set of instances (objects or
phenomena) characterized by a subset of relevant features.
Design models are complete descriptions (features and behaviors) of
targeted artifacts .
[link]
Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Tue, 04/23/2013 - 22:53
thank you for your kind.
---Questions that i should answer---

Q1/What is the purpose of a model?
solution:
the purpose of model is to simplify the representation of
some aspect of the world ,like a formula in Physics to assist
calculations and predictions.
Reference:
[link]

Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Tue, 04/23/2013 - 22:53
> one is modeling the solution design rather than the real world (where
> the problem has not been solved yet). Methodologically, one wants to
> cast the solution design in terms of real world artifacts because
> that will lead to a more robust design. IOW, one designs the software
> solution and then models both the conceptual design and things in the
Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Sun, 04/21/2013 - 00:01
Responding to Mercer-Hursh...

I basically agree, except for a Catch-22. When developing software, one
is modeling the solution design rather than the real world (where the
problem has not been solved yet). Methodologically, one wants to cast
the solution design in terms of real world artifacts because that will

Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Fri, 04/19/2013 - 17:31
Chriss,

Before you ask us this question. You need to answer some questions for us.

First question would be, "What is the purpose of a model?" Look in the
dictionary and tell us!

Next question is "What is a design?" Again look it up!

How would you handle the tension of the need for detail in describing a

Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Fri, 04/19/2013 - 17:31
> diagram or model class diagram?

I would say neither.

One models X (something in the real world)
One designs software (to implement the model of X)
One draws a UML diagram to document the model.
One translates the UML documentation of the model into software to
implement the system.

Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Mon, 04/15/2013 - 22:42
Thank you for your answers
I just wanted to properly describe(in the textual form) my
UML diagrams(use case,class,etc.)
*Which one of these sentences is more appropriate?
design class diagram or model class diagram?
As i conclude from your answer that UML is for modeling as
its name imply,Does this right?
Categories: Communities

Re: [UML Forum] Tiny Types usage

UML Forum Google Group - Mon, 04/15/2013 - 22:42
Responding to Cabral...

Uh, where are the extra classes he is talking about? All I see are user
specified ADTs that will be mapped to primitive hardware types in a
header file for the compiler.

As a translationist, I haven't written any OOPL code in nearly two
decades, but back when I was doing that, this was already SOP and you

Categories: Communities

Re: [UML Forum] N-ary association alternative

UML Forum Google Group - Mon, 04/15/2013 - 22:42
Responding to Cabral...

You don't have to use it just because it is in UML. For example, in the
methodology I use, the /only /relationships used are simple binary and
generalization.

There is always a way to do it with binary relations. Consider the
example used in the UML Reference Manual:

Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Mon, 04/15/2013 - 22:42
I think the OMG spec is obtuse only because it is so complete. I image that
in the beginning there were only class and association. At some point,
duplication became intolerable and the generalization was born. Perhaps the
generalization could have been avoided (100 ways to skin a cat). Premature
Categories: Communities

Re: [UML Forum] understanding class/classifier,modeling/design

UML Forum Google Group - Mon, 04/15/2013 - 22:42
Responding to Chriss...

This sounds like a homework problem. However, since OMG's meta model
documentation is among the most obtuse in the entire technical space,
I'll assume it's not.

Classifier is essentially a very generic element of the UML meta model
that represents some sort of classification of things. Classes and

Categories: Communities

Re: [UML Forum] Data-Centric vs. Object-Oriented

UML Forum Google Group - Mon, 04/15/2013 - 22:42
Respo9nding to George...

So am I -- if one chooses the DDS approach as a design paradigm rather
than a pattern.

--
Life is the only flaw in an otherwise perfect nonexistence
-- Schopenhauer

Imagine how much more difficult physics would be if electrons had feelings
-- Richard Feynman

Categories: Communities

Re: Tiny Types usage

UML Forum Google Group - Mon, 04/15/2013 - 22:42
What do you think about usage of tiny types to modeling your

I don't know why they should be called 'tiny'; to me, coming from an Ada
background, they are just plain types (or perhaps subtypes), and a Very
Good Thing.

I think that the difference between a Class and a Type is whether two

Categories: Communities

Re: [UML Forum] N-ary association alternative

UML Forum Google Group - Mon, 04/15/2013 - 22:42
Replace it with a class.

ɯıɾ-

Sent from my  iPhone

Hi everybody!

Let's suppose in a system, I have a n-ary association among three classes
A, B and C. The problem is the UML tool I'm using (Astah*) doesn't has a
n-ary association symbol.

So, I ask: there is an alternative to a n-ary association with the same

Categories: Communities

Re: [UML Forum] Data-Centric vs. Object-Oriented

UML Forum Google Group - Mon, 04/15/2013 - 12:36
> Fine. I'm just saying that it's not a universal solution.

> You don't
> want to methodologically say, "Oh, I've got distributed
> communications, therefore I must use publish/subscribe."

Note that MOMs typically have an option for direct wiring of components
when that is the indicated type of behavior.

Categories: Communities

Tiny Types usage

UML Forum Google Group - Mon, 04/15/2013 - 12:36
Hi everybody!

What do you think about usage of tiny types to modeling your
systems? Do you agree or disagree? What kind of benefits it
bring to your system?

Some links about this subject:

[link]
[link]

Categories: Communities

N-ary association alternative

UML Forum Google Group - Mon, 04/15/2013 - 12:36
Hi everybody!

Let's suppose in a system, I have a n-ary association among three classes
A, B and C. The problem is the UML tool I'm using (Astah*) doesn't has a
n-ary association symbol.

So, I ask: there is an alternative to a n-ary association with the same
semantics?

Thanks in advance!

Categories: Communities

Re: [UML Forum] Data-Centric vs. Object-Oriented

UML Forum Google Group - Sat, 04/13/2013 - 09:11
"BTW, I would argue DDS is still using messages." - H.S. Lahman

Right. But, let's restrict the definition of "messages" to the outcome of
GENERATE Foo:SomeEvent(parm0, parm1, parm2) TO (someObjectRef); Making
attribute values available in a distributed deployment is a lower-level
concept. Attribute distribution is the job of the DDS domain and doesn't

Categories: Communities