Home » Modeling » UML2 » Behavior and Profiles in metamodeling
| |
Re: Behavior and Profiles in metamodeling [message #472653 is a reply to message #472499] |
Wed, 18 April 2007 22:04 |
Manar El-Kady Messages: 78 Registered: July 2009 |
Member |
|
|
Hi,
Do you mean adding operations like the operations list in the additional
operation section?
In UML specification, the additional operations may be queries or
utility operations that are used to define constraints or other
operations. I don’t know who and when will use these operations (in
metamodeling or modeling )? And what is its impact in the modeling level?
Manar
James Bruck wrote:
> Hi Manar,
>
> some comments below ...
>
> - James.
>
> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> news:euvlf0$lpe$1@build.eclipse.org...
>> Hi,
>>
>> Through the comparison between the different extensions mechanisms, there
>> is a referring to the add/modify behavior. My question is: what does the
>> term behavior mean and how can I decided that my metamodel need to add
>> behavior or not?
>
> The term "behavior" in this case means "operations" .. You might need to
> create additional operations for existing metaclasses .
>
>> Could you illustrate to me how the profiles can’t add behavior, While in
>> UML superstructure (Profiles chapter) tag definitions may be of type
>> operation?
>
> From the spec ... "... a tagged value can only be represented as an
> attribute defined on a steroetype ..."
> The example in the spec defines an attribute of type Operation.
> What the articles are getting at is that there is no facility to actually
> define the behavior of an operation or execute that operation defined on a
> stereotype.
> Section 18.3.8 describes the notion of tag definition and tag value ... "...
> a stereotype may have properties, which may be referred to as tag
> definitions..."
>
>> Moreover, In “Customizing UML” article, it is mentioned that it will be
>> possible to adding behavior to profiles later in the future through the
>> OCL constraints. Is this due to lack of UML specification or due to a
>> limitation in the UML2 tool?
>
> Although constraints could be specified on stereotypes, there is currently
> no way to evaluate those constraints at runtime ( from the opensource API).
> Some tools provide this capability.
>>
>> Thanks in advance,
>> Manar
>
>
|
|
| |
Re: Behavior and Profiles in metamodeling [message #472747 is a reply to message #472707] |
Tue, 24 April 2007 06:42 |
Manar El-Kady Messages: 78 Registered: July 2009 |
Member |
|
|
James,
While creating a profile, there is an operation appeared as one of the
stereotype child. Moreover the (packaged element/ owned Type ) opaque
behavior/Function behavior were appeared as possible children for the
profile element and the stereotype.
Could you illustrate to me how these operations are available to the
profile and the stereotype, while you said that profiles can't add
behavior (which means that it can't add operations).
What is the difference between two elements with the same name but one
is ownedType and the other is packaged element?
Manar
James Bruck wrote:
> Hi Manar,
>
> Some comments below ...
>
> - James.
>
> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> news:f064o1$sva$1@build.eclipse.org...
>> Hi,
>>
>> Do you mean adding operations like the operations list in the additional
>> operation section?
> Yes.
>
>> In UML specification, the additional operations may be queries or
>> utility operations that are used to define constraints or other
>> operations. I don’t know who and when will use these operations (in
>> metamodeling or modeling )?
>
> Those constraints could be added in the metamodel and evaluated at runtime
> when someone is creating an instance of the metamodel ( ie. modeling ).
>
>> And what is its impact in the modeling level?
>
> The constraints could be evaluated to enforce consistency when modeling.
>
>>
>> Manar
>>
>>
>> James Bruck wrote:
>>> Hi Manar,
>>>
>>> some comments below ...
>>>
>>> - James.
>>>
>>> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
>>> news:euvlf0$lpe$1@build.eclipse.org...
>>>> Hi,
>>>>
>>>> Through the comparison between the different extensions mechanisms,
> there
>>>> is a referring to the add/modify behavior. My question is: what does
> the
>>>> term behavior mean and how can I decided that my metamodel need to add
>>>> behavior or not?
>>> The term "behavior" in this case means "operations" .. You might need
> to
>>> create additional operations for existing metaclasses .
>>>
>>>> Could you illustrate to me how the profiles can’t add behavior, While
> in
>>>> UML superstructure (Profiles chapter) tag definitions may be of type
>>>> operation?
>>> From the spec ... "... a tagged value can only be represented as an
>>> attribute defined on a steroetype ..."
>>> The example in the spec defines an attribute of type Operation.
>>> What the articles are getting at is that there is no facility to
> actually
>>> define the behavior of an operation or execute that operation defined on
> a
>>> stereotype.
>>> Section 18.3.8 describes the notion of tag definition and tag value ...
> "...
>>> a stereotype may have properties, which may be referred to as tag
>>> definitions..."
>>>
>>>> Moreover, In “Customizing UML” article, it is mentioned that it will be
>>>> possible to adding behavior to profiles later in the future through the
>>>> OCL constraints. Is this due to lack of UML specification or due to a
>>>> limitation in the UML2 tool?
>>> Although constraints could be specified on stereotypes, there is
> currently
>>> no way to evaluate those constraints at runtime ( from the opensource
> API).
>>> Some tools provide this capability.
>>>> Thanks in advance,
>>>> Manar
>>>
>
>
|
|
|
Re: Behavior and Profiles in metamodeling [message #472799 is a reply to message #472747] |
Wed, 25 April 2007 13:44 |
james bruck Messages: 1724 Registered: July 2009 |
Senior Member |
|
|
Hi Manar,
some comments below...
"Manar El-Kady" <manar.elkady@gmail.com> wrote in message
news:f0k8uq$rrr$1@build.eclipse.org...
> James,
>
> While creating a profile, there is an operation appeared as one of the
> stereotype child. Moreover the (packaged element/ owned Type ) opaque
> behavior/Function behavior were appeared as possible children for the
> profile element and the stereotype.
>
> Could you illustrate to me how these operations are available to the
> profile and the stereotype, while you said that profiles can't add
> behavior (which means that it can't add operations).
Stereotypes extend "Class" which is why you have the inherited ability to
specify operations, however, I believe this is an unintended side effect by
the author trying to reuse exsiting concepts and getting more than what is
really needed. There are many examples of this in the spec including Ports
extending Properties etc. At the end of the day, there is no facility for
evaluating these operations after you apply the stereotype.
>
> What is the difference between two elements with the same name but one
> is ownedType and the other is packaged element?
The collection to which certain elements get added depends on where the spec
indicates they should be added. Packages can have "PackagedElements" for
example.
>
>
> Manar
>
>
>
> James Bruck wrote:
> > Hi Manar,
> >
> > Some comments below ...
> >
> > - James.
> >
> > "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> > news:f064o1$sva$1@build.eclipse.org...
> >> Hi,
> >>
> >> Do you mean adding operations like the operations list in the
additional
> >> operation section?
> > Yes.
> >
> >> In UML specification, the additional operations may be queries or
> >> utility operations that are used to define constraints or other
> >> operations. I don
|
|
|
Re: Behavior and Profiles in metamodeling [message #472902 is a reply to message #472799] |
Mon, 14 May 2007 16:55 |
Kenn Hussey Messages: 1620 Registered: July 2009 |
Senior Member |
|
|
Note that Package::ownedType is a subset of Package::packagedElement, so the
only real difference here (assuming you are creating/adding a type) is the
events that will be fired - adding to the former will fire two events (one
for the ownedType feature and one for the packagedElement feature) whereas
adding to the latter will fire only one event (just for the packagedElement
feature).
Kenn
"James Bruck" <jbruck@ca.ibm.com> wrote in message
news:f0nlvm$dqg$2@build.eclipse.org...
> Hi Manar,
>
> some comments below...
>
> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> news:f0k8uq$rrr$1@build.eclipse.org...
>> James,
>>
>> While creating a profile, there is an operation appeared as one of the
>> stereotype child. Moreover the (packaged element/ owned Type ) opaque
>> behavior/Function behavior were appeared as possible children for the
>> profile element and the stereotype.
>>
>> Could you illustrate to me how these operations are available to the
>> profile and the stereotype, while you said that profiles can't add
>> behavior (which means that it can't add operations).
>
>
> Stereotypes extend "Class" which is why you have the inherited ability to
> specify operations, however, I believe this is an unintended side effect
> by
> the author trying to reuse exsiting concepts and getting more than what is
> really needed. There are many examples of this in the spec including
> Ports
> extending Properties etc. At the end of the day, there is no facility
> for
> evaluating these operations after you apply the stereotype.
>
>
>>
>> What is the difference between two elements with the same name but one
>> is ownedType and the other is packaged element?
>
> The collection to which certain elements get added depends on where the
> spec
> indicates they should be added. Packages can have "PackagedElements" for
> example.
>
>
>>
>>
>> Manar
>>
>>
>>
>> James Bruck wrote:
>> > Hi Manar,
>> >
>> > Some comments below ...
>> >
>> > - James.
>> >
>> > "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
>> > news:f064o1$sva$1@build.eclipse.org...
>> >> Hi,
>> >>
>> >> Do you mean adding operations like the operations list in the
> additional
>> >> operation section?
>> > Yes.
>> >
>> >> In UML specification, the additional operations may be queries or
>> >> utility operations that are used to define constraints or other
>> >> operations. I don't know who and when will use these operations (in
>> >> metamodeling or modeling )?
>> >
>> > Those constraints could be added in the metamodel and evaluated at
> runtime
>> > when someone is creating an instance of the metamodel ( ie. modeling ).
>> >
>> >> And what is its impact in the modeling level?
>> >
>> > The constraints could be evaluated to enforce consistency when
>> > modeling.
>> >
>> >>
>> >> Manar
>> >>
>> >>
>> >> James Bruck wrote:
>> >>> Hi Manar,
>> >>>
>> >>> some comments below ...
>> >>>
>> >>> - James.
>> >>>
>> >>> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
>> >>> news:euvlf0$lpe$1@build.eclipse.org...
>> >>>> Hi,
>> >>>>
>> >>>> Through the comparison between the different extensions mechanisms,
>> > there
>> >>>> is a referring to the add/modify behavior. My question is: what does
>> > the
>> >>>> term behavior mean and how can I decided that my metamodel need to
> add
>> >>>> behavior or not?
>> >>> The term "behavior" in this case means "operations" .. You might
> need
>> > to
>> >>> create additional operations for existing metaclasses .
>> >>>
>> >>>> Could you illustrate to me how the profiles can't add behavior,
>> >>>> While
>> > in
>> >>>> UML superstructure (Profiles chapter) tag definitions may be of type
>> >>>> operation?
>> >>> From the spec ... "... a tagged value can only be represented as an
>> >>> attribute defined on a steroetype ..."
>> >>> The example in the spec defines an attribute of type Operation.
>> >>> What the articles are getting at is that there is no facility to
>> > actually
>> >>> define the behavior of an operation or execute that operation defined
> on
>> > a
>> >>> stereotype.
>> >>> Section 18.3.8 describes the notion of tag definition and tag value
> ...
>> > "...
>> >>> a stereotype may have properties, which may be referred to as tag
>> >>> definitions..."
>> >>>
>> >>>> Moreover, In "Customizing UML" article, it is mentioned that it will
> be
>> >>>> possible to adding behavior to profiles later in the future through
> the
>> >>>> OCL constraints. Is this due to lack of UML specification or due to
>> >>>> a
>> >>>> limitation in the UML2 tool?
>> >>> Although constraints could be specified on stereotypes, there is
>> > currently
>> >>> no way to evaluate those constraints at runtime ( from the opensource
>> > API).
>> >>> Some tools provide this capability.
>> >>>> Thanks in advance,
>> >>>> Manar
>> >>>
>> >
>> >
>
>
|
|
| |
Re: Behavior and Profiles in metamodeling [message #607716 is a reply to message #472499] |
Wed, 18 April 2007 22:04 |
Manar El-Kady Messages: 78 Registered: July 2009 |
Member |
|
|
Hi,
Do you mean adding operations like the operations list in the additional
operation section?
In UML specification, the additional operations may be queries or
utility operations that are used to define constraints or other
operations. I don’t know who and when will use these operations (in
metamodeling or modeling )? And what is its impact in the modeling level?
Manar
James Bruck wrote:
> Hi Manar,
>
> some comments below ...
>
> - James.
>
> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> news:euvlf0$lpe$1@build.eclipse.org...
>> Hi,
>>
>> Through the comparison between the different extensions mechanisms, there
>> is a referring to the add/modify behavior. My question is: what does the
>> term behavior mean and how can I decided that my metamodel need to add
>> behavior or not?
>
> The term "behavior" in this case means "operations" .. You might need to
> create additional operations for existing metaclasses .
>
>> Could you illustrate to me how the profiles can’t add behavior, While in
>> UML superstructure (Profiles chapter) tag definitions may be of type
>> operation?
>
> From the spec ... "... a tagged value can only be represented as an
> attribute defined on a steroetype ..."
> The example in the spec defines an attribute of type Operation.
> What the articles are getting at is that there is no facility to actually
> define the behavior of an operation or execute that operation defined on a
> stereotype.
> Section 18.3.8 describes the notion of tag definition and tag value ... "...
> a stereotype may have properties, which may be referred to as tag
> definitions..."
>
>> Moreover, In “Customizing UML” article, it is mentioned that it will be
>> possible to adding behavior to profiles later in the future through the
>> OCL constraints. Is this due to lack of UML specification or due to a
>> limitation in the UML2 tool?
>
> Although constraints could be specified on stereotypes, there is currently
> no way to evaluate those constraints at runtime ( from the opensource API).
> Some tools provide this capability.
>>
>> Thanks in advance,
>> Manar
>
>
|
|
| |
Re: Behavior and Profiles in metamodeling [message #611042 is a reply to message #472707] |
Tue, 24 April 2007 06:42 |
Manar El-Kady Messages: 78 Registered: July 2009 |
Member |
|
|
James,
While creating a profile, there is an operation appeared as one of the
stereotype child. Moreover the (packaged element/ owned Type ) opaque
behavior/Function behavior were appeared as possible children for the
profile element and the stereotype.
Could you illustrate to me how these operations are available to the
profile and the stereotype, while you said that profiles can't add
behavior (which means that it can't add operations).
What is the difference between two elements with the same name but one
is ownedType and the other is packaged element?
Manar
James Bruck wrote:
> Hi Manar,
>
> Some comments below ...
>
> - James.
>
> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> news:f064o1$sva$1@build.eclipse.org...
>> Hi,
>>
>> Do you mean adding operations like the operations list in the additional
>> operation section?
> Yes.
>
>> In UML specification, the additional operations may be queries or
>> utility operations that are used to define constraints or other
>> operations. I don’t know who and when will use these operations (in
>> metamodeling or modeling )?
>
> Those constraints could be added in the metamodel and evaluated at runtime
> when someone is creating an instance of the metamodel ( ie. modeling ).
>
>> And what is its impact in the modeling level?
>
> The constraints could be evaluated to enforce consistency when modeling.
>
>>
>> Manar
>>
>>
>> James Bruck wrote:
>>> Hi Manar,
>>>
>>> some comments below ...
>>>
>>> - James.
>>>
>>> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
>>> news:euvlf0$lpe$1@build.eclipse.org...
>>>> Hi,
>>>>
>>>> Through the comparison between the different extensions mechanisms,
> there
>>>> is a referring to the add/modify behavior. My question is: what does
> the
>>>> term behavior mean and how can I decided that my metamodel need to add
>>>> behavior or not?
>>> The term "behavior" in this case means "operations" .. You might need
> to
>>> create additional operations for existing metaclasses .
>>>
>>>> Could you illustrate to me how the profiles can’t add behavior, While
> in
>>>> UML superstructure (Profiles chapter) tag definitions may be of type
>>>> operation?
>>> From the spec ... "... a tagged value can only be represented as an
>>> attribute defined on a steroetype ..."
>>> The example in the spec defines an attribute of type Operation.
>>> What the articles are getting at is that there is no facility to
> actually
>>> define the behavior of an operation or execute that operation defined on
> a
>>> stereotype.
>>> Section 18.3.8 describes the notion of tag definition and tag value ...
> "...
>>> a stereotype may have properties, which may be referred to as tag
>>> definitions..."
>>>
>>>> Moreover, In “Customizing UML” article, it is mentioned that it will be
>>>> possible to adding behavior to profiles later in the future through the
>>>> OCL constraints. Is this due to lack of UML specification or due to a
>>>> limitation in the UML2 tool?
>>> Although constraints could be specified on stereotypes, there is
> currently
>>> no way to evaluate those constraints at runtime ( from the opensource
> API).
>>> Some tools provide this capability.
>>>> Thanks in advance,
>>>> Manar
>>>
>
>
|
|
|
Re: Behavior and Profiles in metamodeling [message #613181 is a reply to message #472747] |
Wed, 25 April 2007 13:44 |
james bruck Messages: 1724 Registered: July 2009 |
Senior Member |
|
|
Hi Manar,
some comments below...
"Manar El-Kady" <manar.elkady@gmail.com> wrote in message
news:f0k8uq$rrr$1@build.eclipse.org...
> James,
>
> While creating a profile, there is an operation appeared as one of the
> stereotype child. Moreover the (packaged element/ owned Type ) opaque
> behavior/Function behavior were appeared as possible children for the
> profile element and the stereotype.
>
> Could you illustrate to me how these operations are available to the
> profile and the stereotype, while you said that profiles can't add
> behavior (which means that it can't add operations).
Stereotypes extend "Class" which is why you have the inherited ability to
specify operations, however, I believe this is an unintended side effect by
the author trying to reuse exsiting concepts and getting more than what is
really needed. There are many examples of this in the spec including Ports
extending Properties etc. At the end of the day, there is no facility for
evaluating these operations after you apply the stereotype.
>
> What is the difference between two elements with the same name but one
> is ownedType and the other is packaged element?
The collection to which certain elements get added depends on where the spec
indicates they should be added. Packages can have "PackagedElements" for
example.
>
>
> Manar
>
>
>
> James Bruck wrote:
> > Hi Manar,
> >
> > Some comments below ...
> >
> > - James.
> >
> > "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> > news:f064o1$sva$1@build.eclipse.org...
> >> Hi,
> >>
> >> Do you mean adding operations like the operations list in the
additional
> >> operation section?
> > Yes.
> >
> >> In UML specification, the additional operations may be queries or
> >> utility operations that are used to define constraints or other
> >> operations. I don
|
|
|
Re: Behavior and Profiles in metamodeling [message #615859 is a reply to message #472799] |
Mon, 14 May 2007 16:55 |
Kenn Hussey Messages: 1620 Registered: July 2009 |
Senior Member |
|
|
Note that Package::ownedType is a subset of Package::packagedElement, so the
only real difference here (assuming you are creating/adding a type) is the
events that will be fired - adding to the former will fire two events (one
for the ownedType feature and one for the packagedElement feature) whereas
adding to the latter will fire only one event (just for the packagedElement
feature).
Kenn
"James Bruck" <jbruck@ca.ibm.com> wrote in message
news:f0nlvm$dqg$2@build.eclipse.org...
> Hi Manar,
>
> some comments below...
>
> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
> news:f0k8uq$rrr$1@build.eclipse.org...
>> James,
>>
>> While creating a profile, there is an operation appeared as one of the
>> stereotype child. Moreover the (packaged element/ owned Type ) opaque
>> behavior/Function behavior were appeared as possible children for the
>> profile element and the stereotype.
>>
>> Could you illustrate to me how these operations are available to the
>> profile and the stereotype, while you said that profiles can't add
>> behavior (which means that it can't add operations).
>
>
> Stereotypes extend "Class" which is why you have the inherited ability to
> specify operations, however, I believe this is an unintended side effect
> by
> the author trying to reuse exsiting concepts and getting more than what is
> really needed. There are many examples of this in the spec including
> Ports
> extending Properties etc. At the end of the day, there is no facility
> for
> evaluating these operations after you apply the stereotype.
>
>
>>
>> What is the difference between two elements with the same name but one
>> is ownedType and the other is packaged element?
>
> The collection to which certain elements get added depends on where the
> spec
> indicates they should be added. Packages can have "PackagedElements" for
> example.
>
>
>>
>>
>> Manar
>>
>>
>>
>> James Bruck wrote:
>> > Hi Manar,
>> >
>> > Some comments below ...
>> >
>> > - James.
>> >
>> > "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
>> > news:f064o1$sva$1@build.eclipse.org...
>> >> Hi,
>> >>
>> >> Do you mean adding operations like the operations list in the
> additional
>> >> operation section?
>> > Yes.
>> >
>> >> In UML specification, the additional operations may be queries or
>> >> utility operations that are used to define constraints or other
>> >> operations. I don't know who and when will use these operations (in
>> >> metamodeling or modeling )?
>> >
>> > Those constraints could be added in the metamodel and evaluated at
> runtime
>> > when someone is creating an instance of the metamodel ( ie. modeling ).
>> >
>> >> And what is its impact in the modeling level?
>> >
>> > The constraints could be evaluated to enforce consistency when
>> > modeling.
>> >
>> >>
>> >> Manar
>> >>
>> >>
>> >> James Bruck wrote:
>> >>> Hi Manar,
>> >>>
>> >>> some comments below ...
>> >>>
>> >>> - James.
>> >>>
>> >>> "Manar El-Kady" <manar.elkady@gmail.com> wrote in message
>> >>> news:euvlf0$lpe$1@build.eclipse.org...
>> >>>> Hi,
>> >>>>
>> >>>> Through the comparison between the different extensions mechanisms,
>> > there
>> >>>> is a referring to the add/modify behavior. My question is: what does
>> > the
>> >>>> term behavior mean and how can I decided that my metamodel need to
> add
>> >>>> behavior or not?
>> >>> The term "behavior" in this case means "operations" .. You might
> need
>> > to
>> >>> create additional operations for existing metaclasses .
>> >>>
>> >>>> Could you illustrate to me how the profiles can't add behavior,
>> >>>> While
>> > in
>> >>>> UML superstructure (Profiles chapter) tag definitions may be of type
>> >>>> operation?
>> >>> From the spec ... "... a tagged value can only be represented as an
>> >>> attribute defined on a steroetype ..."
>> >>> The example in the spec defines an attribute of type Operation.
>> >>> What the articles are getting at is that there is no facility to
>> > actually
>> >>> define the behavior of an operation or execute that operation defined
> on
>> > a
>> >>> stereotype.
>> >>> Section 18.3.8 describes the notion of tag definition and tag value
> ...
>> > "...
>> >>> a stereotype may have properties, which may be referred to as tag
>> >>> definitions..."
>> >>>
>> >>>> Moreover, In "Customizing UML" article, it is mentioned that it will
> be
>> >>>> possible to adding behavior to profiles later in the future through
> the
>> >>>> OCL constraints. Is this due to lack of UML specification or due to
>> >>>> a
>> >>>> limitation in the UML2 tool?
>> >>> Although constraints could be specified on stereotypes, there is
>> > currently
>> >>> no way to evaluate those constraints at runtime ( from the opensource
>> > API).
>> >>> Some tools provide this capability.
>> >>>> Thanks in advance,
>> >>>> Manar
>> >>>
>> >
>> >
>
>
|
|
|
Goto Forum:
Current Time: Wed Sep 25 02:23:16 GMT 2024
Powered by FUDForum. Page generated in 0.03497 seconds
|