Home » Modeling » UML2 » Is it reasonable to want EDataType stereotype to be applyable to DataType
| | | |
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #472813 is a reply to message #472812] |
Thu, 26 April 2007 17:19 |
Tas Frangoullides Messages: 195 Registered: July 2009 |
Senior Member |
|
|
Hi Kenn,
It would help becuase Instance Class Name is what I am after, but on the
profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
2.0.2.. has this changed since then?
Tas
"Kenn Hussey" <khussey@ca.ibm.com> wrote in message
news:f0qmcp$ace$1@build.eclipse.org...
> Tas,
>
> Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
> think it makes sense for the EDataType stereotype extend only
> PrimitiveType. Note that you can stereotype a UML data type as <<eClass>>
> if you want to specify some of the tags that are available from the
> EDataType stereotype...
>
> Kenn
>
> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> message news:f0qfvo$ug9$1@build.eclipse.org...
>> Hi James,
>>
>> I believe I understand the thinking there and it may be one of those
>> highly subjective areas; What is considered primtive may depend on what's
>> being modeled.
>>
>> Some other itneresting poitns are:
>> * A PrimitiveType can have properties and operations
>> * A Date (and many other standard java classes) have substructure yet
>> you need to model them as primitves to use UML2Ecore.
>>
>> Do you think it would be contentious to allow the EDataType Stereotype to
>> be applicable to both DataType and PrimitiveType?
>>
>> Thanks,
>> Tas
>>
>>
>> "James Bruck" <jbruck@ca.ibm.com> wrote in message
>> news:f0qcvr$567$2@build.eclipse.org...
>>> Hi Tas,
>>>
>>> That's a good question ... there definitely was some back and forth when
>>> this decision was made so it was not a clear cut decision.
>>> At one point the XMLPrimitiveTypes library used UML DataTypes as opposed
>>> to
>>> Primitive types.
>>>
>>> The key point is that is that PrimitiveTypes are DataTypes but ones
>>> that
>>> don't have substructure ( see section 7.3.43 ).
>>>
>>> This maps more closely to EDataTypes in ecore.
>>>
>>> Regards,
>>>
>>> - James.
>>>
>>>
>>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>>> message
>>> news:f0pr0n$las$1@build.eclipse.org...
>>>> Hi all,
>>>>
>>>> Is there a partciular reason that EDataType cannot be applied to a UML
>>>> DataType. I often use DataType in my models becuase it feels like the
>>> right
>>>> choice at that level of abstraction. If I need to process models
>>>> through
>>>> UML2Ecore I use my own profile and stereotype and trasnform the model
>>> first,
>>>> replacing DataType with Primitive Type + EDataType stereotype.. This is
>>> fine
>>>> but it would *nice* if I didn't have to do it.
>>>>
>>>> Tas.
>>>>
>>>>
>>>
>>>
>>
>>
>
>
|
|
|
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #472815 is a reply to message #472813] |
Thu, 26 April 2007 18:59 |
james bruck Messages: 1724 Registered: July 2009 |
Senior Member |
|
|
Hi Tas,
There have been some recent changes in that area, you might want to have a
look at 2.1.
You can raise a bugzilla on that if it doesn't behave the way you expect in
2.1. Please add a detaile description of the issue you are trying to solve.
- James.
"Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in message
news:f0qn0t$bme$1@build.eclipse.org...
> Hi Kenn,
>
> It would help becuase Instance Class Name is what I am after, but on the
> profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
> 2.0.2.. has this changed since then?
>
> Tas
>
>
> "Kenn Hussey" <khussey@ca.ibm.com> wrote in message
> news:f0qmcp$ace$1@build.eclipse.org...
> > Tas,
> >
> > Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
> > think it makes sense for the EDataType stereotype extend only
> > PrimitiveType. Note that you can stereotype a UML data type as
<<eClass>>
> > if you want to specify some of the tags that are available from the
> > EDataType stereotype...
> >
> > Kenn
> >
> > "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> > message news:f0qfvo$ug9$1@build.eclipse.org...
> >> Hi James,
> >>
> >> I believe I understand the thinking there and it may be one of those
> >> highly subjective areas; What is considered primtive may depend on
what's
> >> being modeled.
> >>
> >> Some other itneresting poitns are:
> >> * A PrimitiveType can have properties and operations
> >> * A Date (and many other standard java classes) have substructure yet
> >> you need to model them as primitves to use UML2Ecore.
> >>
> >> Do you think it would be contentious to allow the EDataType Stereotype
to
> >> be applicable to both DataType and PrimitiveType?
> >>
> >> Thanks,
> >> Tas
> >>
> >>
> >> "James Bruck" <jbruck@ca.ibm.com> wrote in message
> >> news:f0qcvr$567$2@build.eclipse.org...
> >>> Hi Tas,
> >>>
> >>> That's a good question ... there definitely was some back and forth
when
> >>> this decision was made so it was not a clear cut decision.
> >>> At one point the XMLPrimitiveTypes library used UML DataTypes as
opposed
> >>> to
> >>> Primitive types.
> >>>
> >>> The key point is that is that PrimitiveTypes are DataTypes but ones
> >>> that
> >>> don't have substructure ( see section 7.3.43 ).
> >>>
> >>> This maps more closely to EDataTypes in ecore.
> >>>
> >>> Regards,
> >>>
> >>> - James.
> >>>
> >>>
> >>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> >>> message
> >>> news:f0pr0n$las$1@build.eclipse.org...
> >>>> Hi all,
> >>>>
> >>>> Is there a partciular reason that EDataType cannot be applied to a
UML
> >>>> DataType. I often use DataType in my models becuase it feels like the
> >>> right
> >>>> choice at that level of abstraction. If I need to process models
> >>>> through
> >>>> UML2Ecore I use my own profile and stereotype and trasnform the model
> >>> first,
> >>>> replacing DataType with Primitive Type + EDataType stereotype.. This
is
> >>> fine
> >>>> but it would *nice* if I didn't have to do it.
> >>>>
> >>>> Tas.
> >>>>
> >>>>
> >>>
> >>>
> >>
> >>
> >
> >
>
>
|
|
|
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #472817 is a reply to message #472815] |
Thu, 26 April 2007 21:02 |
Tas Frangoullides Messages: 195 Registered: July 2009 |
Senior Member |
|
|
James,
I had a feeling you were going to say that :-)
I'm a bit stuck actually becuase we are deploying a solution built on top of
RSM 7.0 which uses 2.02. Unless a change has been made since then... time
for me to check the fix packs!
Thanks for your help,
Tas
"James Bruck" <jbruck@ca.ibm.com> wrote in message
news:f0qsq5$tnl$2@build.eclipse.org...
> Hi Tas,
>
> There have been some recent changes in that area, you might want to have
> a
> look at 2.1.
> You can raise a bugzilla on that if it doesn't behave the way you expect
> in
> 2.1. Please add a detaile description of the issue you are trying to
> solve.
>
> - James.
>
>
> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> message
> news:f0qn0t$bme$1@build.eclipse.org...
>> Hi Kenn,
>>
>> It would help becuase Instance Class Name is what I am after, but on the
>> profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
>> 2.0.2.. has this changed since then?
>>
>> Tas
>>
>>
>> "Kenn Hussey" <khussey@ca.ibm.com> wrote in message
>> news:f0qmcp$ace$1@build.eclipse.org...
>> > Tas,
>> >
>> > Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
>> > think it makes sense for the EDataType stereotype extend only
>> > PrimitiveType. Note that you can stereotype a UML data type as
> <<eClass>>
>> > if you want to specify some of the tags that are available from the
>> > EDataType stereotype...
>> >
>> > Kenn
>> >
>> > "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>> > message news:f0qfvo$ug9$1@build.eclipse.org...
>> >> Hi James,
>> >>
>> >> I believe I understand the thinking there and it may be one of those
>> >> highly subjective areas; What is considered primtive may depend on
> what's
>> >> being modeled.
>> >>
>> >> Some other itneresting poitns are:
>> >> * A PrimitiveType can have properties and operations
>> >> * A Date (and many other standard java classes) have substructure
>> >> yet
>> >> you need to model them as primitves to use UML2Ecore.
>> >>
>> >> Do you think it would be contentious to allow the EDataType Stereotype
> to
>> >> be applicable to both DataType and PrimitiveType?
>> >>
>> >> Thanks,
>> >> Tas
>> >>
>> >>
>> >> "James Bruck" <jbruck@ca.ibm.com> wrote in message
>> >> news:f0qcvr$567$2@build.eclipse.org...
>> >>> Hi Tas,
>> >>>
>> >>> That's a good question ... there definitely was some back and forth
> when
>> >>> this decision was made so it was not a clear cut decision.
>> >>> At one point the XMLPrimitiveTypes library used UML DataTypes as
> opposed
>> >>> to
>> >>> Primitive types.
>> >>>
>> >>> The key point is that is that PrimitiveTypes are DataTypes but ones
>> >>> that
>> >>> don't have substructure ( see section 7.3.43 ).
>> >>>
>> >>> This maps more closely to EDataTypes in ecore.
>> >>>
>> >>> Regards,
>> >>>
>> >>> - James.
>> >>>
>> >>>
>> >>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>> >>> message
>> >>> news:f0pr0n$las$1@build.eclipse.org...
>> >>>> Hi all,
>> >>>>
>> >>>> Is there a partciular reason that EDataType cannot be applied to a
> UML
>> >>>> DataType. I often use DataType in my models becuase it feels like
>> >>>> the
>> >>> right
>> >>>> choice at that level of abstraction. If I need to process models
>> >>>> through
>> >>>> UML2Ecore I use my own profile and stereotype and trasnform the
>> >>>> model
>> >>> first,
>> >>>> replacing DataType with Primitive Type + EDataType stereotype.. This
> is
>> >>> fine
>> >>>> but it would *nice* if I didn't have to do it.
>> >>>>
>> >>>> Tas.
>> >>>>
>> >>>>
>> >>>
>> >>>
>> >>
>> >>
>> >
>> >
>>
>>
>
>
|
|
|
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #472818 is a reply to message #472813] |
Thu, 26 April 2007 22:03 |
Kenn Hussey Messages: 1620 Registered: July 2009 |
Senior Member |
|
|
Tas,
The EDataType stereotype extends DataType as of UML2 2.1.
Kenn
"Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in message
news:f0qn0t$bme$1@build.eclipse.org...
> Hi Kenn,
>
> It would help becuase Instance Class Name is what I am after, but on the
> profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
> 2.0.2.. has this changed since then?
>
> Tas
>
>
> "Kenn Hussey" <khussey@ca.ibm.com> wrote in message
> news:f0qmcp$ace$1@build.eclipse.org...
>> Tas,
>>
>> Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
>> think it makes sense for the EDataType stereotype extend only
>> PrimitiveType. Note that you can stereotype a UML data type as <<eClass>>
>> if you want to specify some of the tags that are available from the
>> EDataType stereotype...
>>
>> Kenn
>>
>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>> message news:f0qfvo$ug9$1@build.eclipse.org...
>>> Hi James,
>>>
>>> I believe I understand the thinking there and it may be one of those
>>> highly subjective areas; What is considered primtive may depend on
>>> what's being modeled.
>>>
>>> Some other itneresting poitns are:
>>> * A PrimitiveType can have properties and operations
>>> * A Date (and many other standard java classes) have substructure yet
>>> you need to model them as primitves to use UML2Ecore.
>>>
>>> Do you think it would be contentious to allow the EDataType Stereotype
>>> to be applicable to both DataType and PrimitiveType?
>>>
>>> Thanks,
>>> Tas
>>>
>>>
>>> "James Bruck" <jbruck@ca.ibm.com> wrote in message
>>> news:f0qcvr$567$2@build.eclipse.org...
>>>> Hi Tas,
>>>>
>>>> That's a good question ... there definitely was some back and forth
>>>> when
>>>> this decision was made so it was not a clear cut decision.
>>>> At one point the XMLPrimitiveTypes library used UML DataTypes as
>>>> opposed to
>>>> Primitive types.
>>>>
>>>> The key point is that is that PrimitiveTypes are DataTypes but ones
>>>> that
>>>> don't have substructure ( see section 7.3.43 ).
>>>>
>>>> This maps more closely to EDataTypes in ecore.
>>>>
>>>> Regards,
>>>>
>>>> - James.
>>>>
>>>>
>>>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>>>> message
>>>> news:f0pr0n$las$1@build.eclipse.org...
>>>>> Hi all,
>>>>>
>>>>> Is there a partciular reason that EDataType cannot be applied to a UML
>>>>> DataType. I often use DataType in my models becuase it feels like the
>>>> right
>>>>> choice at that level of abstraction. If I need to process models
>>>>> through
>>>>> UML2Ecore I use my own profile and stereotype and trasnform the model
>>>> first,
>>>>> replacing DataType with Primitive Type + EDataType stereotype.. This
>>>>> is
>>>> fine
>>>>> but it would *nice* if I didn't have to do it.
>>>>>
>>>>> Tas.
>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>
>
|
|
| | | |
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #613195 is a reply to message #472812] |
Thu, 26 April 2007 17:19 |
Tas Frangoullides Messages: 195 Registered: July 2009 |
Senior Member |
|
|
Hi Kenn,
It would help becuase Instance Class Name is what I am after, but on the
profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
2.0.2.. has this changed since then?
Tas
"Kenn Hussey" <khussey@ca.ibm.com> wrote in message
news:f0qmcp$ace$1@build.eclipse.org...
> Tas,
>
> Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
> think it makes sense for the EDataType stereotype extend only
> PrimitiveType. Note that you can stereotype a UML data type as <<eClass>>
> if you want to specify some of the tags that are available from the
> EDataType stereotype...
>
> Kenn
>
> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> message news:f0qfvo$ug9$1@build.eclipse.org...
>> Hi James,
>>
>> I believe I understand the thinking there and it may be one of those
>> highly subjective areas; What is considered primtive may depend on what's
>> being modeled.
>>
>> Some other itneresting poitns are:
>> * A PrimitiveType can have properties and operations
>> * A Date (and many other standard java classes) have substructure yet
>> you need to model them as primitves to use UML2Ecore.
>>
>> Do you think it would be contentious to allow the EDataType Stereotype to
>> be applicable to both DataType and PrimitiveType?
>>
>> Thanks,
>> Tas
>>
>>
>> "James Bruck" <jbruck@ca.ibm.com> wrote in message
>> news:f0qcvr$567$2@build.eclipse.org...
>>> Hi Tas,
>>>
>>> That's a good question ... there definitely was some back and forth when
>>> this decision was made so it was not a clear cut decision.
>>> At one point the XMLPrimitiveTypes library used UML DataTypes as opposed
>>> to
>>> Primitive types.
>>>
>>> The key point is that is that PrimitiveTypes are DataTypes but ones
>>> that
>>> don't have substructure ( see section 7.3.43 ).
>>>
>>> This maps more closely to EDataTypes in ecore.
>>>
>>> Regards,
>>>
>>> - James.
>>>
>>>
>>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>>> message
>>> news:f0pr0n$las$1@build.eclipse.org...
>>>> Hi all,
>>>>
>>>> Is there a partciular reason that EDataType cannot be applied to a UML
>>>> DataType. I often use DataType in my models becuase it feels like the
>>> right
>>>> choice at that level of abstraction. If I need to process models
>>>> through
>>>> UML2Ecore I use my own profile and stereotype and trasnform the model
>>> first,
>>>> replacing DataType with Primitive Type + EDataType stereotype.. This is
>>> fine
>>>> but it would *nice* if I didn't have to do it.
>>>>
>>>> Tas.
>>>>
>>>>
>>>
>>>
>>
>>
>
>
|
|
|
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #613197 is a reply to message #472813] |
Thu, 26 April 2007 18:59 |
james bruck Messages: 1724 Registered: July 2009 |
Senior Member |
|
|
Hi Tas,
There have been some recent changes in that area, you might want to have a
look at 2.1.
You can raise a bugzilla on that if it doesn't behave the way you expect in
2.1. Please add a detaile description of the issue you are trying to solve.
- James.
"Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in message
news:f0qn0t$bme$1@build.eclipse.org...
> Hi Kenn,
>
> It would help becuase Instance Class Name is what I am after, but on the
> profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
> 2.0.2.. has this changed since then?
>
> Tas
>
>
> "Kenn Hussey" <khussey@ca.ibm.com> wrote in message
> news:f0qmcp$ace$1@build.eclipse.org...
> > Tas,
> >
> > Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
> > think it makes sense for the EDataType stereotype extend only
> > PrimitiveType. Note that you can stereotype a UML data type as
<<eClass>>
> > if you want to specify some of the tags that are available from the
> > EDataType stereotype...
> >
> > Kenn
> >
> > "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> > message news:f0qfvo$ug9$1@build.eclipse.org...
> >> Hi James,
> >>
> >> I believe I understand the thinking there and it may be one of those
> >> highly subjective areas; What is considered primtive may depend on
what's
> >> being modeled.
> >>
> >> Some other itneresting poitns are:
> >> * A PrimitiveType can have properties and operations
> >> * A Date (and many other standard java classes) have substructure yet
> >> you need to model them as primitves to use UML2Ecore.
> >>
> >> Do you think it would be contentious to allow the EDataType Stereotype
to
> >> be applicable to both DataType and PrimitiveType?
> >>
> >> Thanks,
> >> Tas
> >>
> >>
> >> "James Bruck" <jbruck@ca.ibm.com> wrote in message
> >> news:f0qcvr$567$2@build.eclipse.org...
> >>> Hi Tas,
> >>>
> >>> That's a good question ... there definitely was some back and forth
when
> >>> this decision was made so it was not a clear cut decision.
> >>> At one point the XMLPrimitiveTypes library used UML DataTypes as
opposed
> >>> to
> >>> Primitive types.
> >>>
> >>> The key point is that is that PrimitiveTypes are DataTypes but ones
> >>> that
> >>> don't have substructure ( see section 7.3.43 ).
> >>>
> >>> This maps more closely to EDataTypes in ecore.
> >>>
> >>> Regards,
> >>>
> >>> - James.
> >>>
> >>>
> >>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> >>> message
> >>> news:f0pr0n$las$1@build.eclipse.org...
> >>>> Hi all,
> >>>>
> >>>> Is there a partciular reason that EDataType cannot be applied to a
UML
> >>>> DataType. I often use DataType in my models becuase it feels like the
> >>> right
> >>>> choice at that level of abstraction. If I need to process models
> >>>> through
> >>>> UML2Ecore I use my own profile and stereotype and trasnform the model
> >>> first,
> >>>> replacing DataType with Primitive Type + EDataType stereotype.. This
is
> >>> fine
> >>>> but it would *nice* if I didn't have to do it.
> >>>>
> >>>> Tas.
> >>>>
> >>>>
> >>>
> >>>
> >>
> >>
> >
> >
>
>
|
|
|
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #613199 is a reply to message #472815] |
Thu, 26 April 2007 21:02 |
Tas Frangoullides Messages: 195 Registered: July 2009 |
Senior Member |
|
|
James,
I had a feeling you were going to say that :-)
I'm a bit stuck actually becuase we are deploying a solution built on top of
RSM 7.0 which uses 2.02. Unless a change has been made since then... time
for me to check the fix packs!
Thanks for your help,
Tas
"James Bruck" <jbruck@ca.ibm.com> wrote in message
news:f0qsq5$tnl$2@build.eclipse.org...
> Hi Tas,
>
> There have been some recent changes in that area, you might want to have
> a
> look at 2.1.
> You can raise a bugzilla on that if it doesn't behave the way you expect
> in
> 2.1. Please add a detaile description of the issue you are trying to
> solve.
>
> - James.
>
>
> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
> message
> news:f0qn0t$bme$1@build.eclipse.org...
>> Hi Kenn,
>>
>> It would help becuase Instance Class Name is what I am after, but on the
>> profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
>> 2.0.2.. has this changed since then?
>>
>> Tas
>>
>>
>> "Kenn Hussey" <khussey@ca.ibm.com> wrote in message
>> news:f0qmcp$ace$1@build.eclipse.org...
>> > Tas,
>> >
>> > Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
>> > think it makes sense for the EDataType stereotype extend only
>> > PrimitiveType. Note that you can stereotype a UML data type as
> <<eClass>>
>> > if you want to specify some of the tags that are available from the
>> > EDataType stereotype...
>> >
>> > Kenn
>> >
>> > "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>> > message news:f0qfvo$ug9$1@build.eclipse.org...
>> >> Hi James,
>> >>
>> >> I believe I understand the thinking there and it may be one of those
>> >> highly subjective areas; What is considered primtive may depend on
> what's
>> >> being modeled.
>> >>
>> >> Some other itneresting poitns are:
>> >> * A PrimitiveType can have properties and operations
>> >> * A Date (and many other standard java classes) have substructure
>> >> yet
>> >> you need to model them as primitves to use UML2Ecore.
>> >>
>> >> Do you think it would be contentious to allow the EDataType Stereotype
> to
>> >> be applicable to both DataType and PrimitiveType?
>> >>
>> >> Thanks,
>> >> Tas
>> >>
>> >>
>> >> "James Bruck" <jbruck@ca.ibm.com> wrote in message
>> >> news:f0qcvr$567$2@build.eclipse.org...
>> >>> Hi Tas,
>> >>>
>> >>> That's a good question ... there definitely was some back and forth
> when
>> >>> this decision was made so it was not a clear cut decision.
>> >>> At one point the XMLPrimitiveTypes library used UML DataTypes as
> opposed
>> >>> to
>> >>> Primitive types.
>> >>>
>> >>> The key point is that is that PrimitiveTypes are DataTypes but ones
>> >>> that
>> >>> don't have substructure ( see section 7.3.43 ).
>> >>>
>> >>> This maps more closely to EDataTypes in ecore.
>> >>>
>> >>> Regards,
>> >>>
>> >>> - James.
>> >>>
>> >>>
>> >>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>> >>> message
>> >>> news:f0pr0n$las$1@build.eclipse.org...
>> >>>> Hi all,
>> >>>>
>> >>>> Is there a partciular reason that EDataType cannot be applied to a
> UML
>> >>>> DataType. I often use DataType in my models becuase it feels like
>> >>>> the
>> >>> right
>> >>>> choice at that level of abstraction. If I need to process models
>> >>>> through
>> >>>> UML2Ecore I use my own profile and stereotype and trasnform the
>> >>>> model
>> >>> first,
>> >>>> replacing DataType with Primitive Type + EDataType stereotype.. This
> is
>> >>> fine
>> >>>> but it would *nice* if I didn't have to do it.
>> >>>>
>> >>>> Tas.
>> >>>>
>> >>>>
>> >>>
>> >>>
>> >>
>> >>
>> >
>> >
>>
>>
>
>
|
|
|
Re: Is it reasonable to want EDataType stereotype to be applyable to DataType [message #613736 is a reply to message #472813] |
Thu, 26 April 2007 22:03 |
Kenn Hussey Messages: 1620 Registered: July 2009 |
Senior Member |
|
|
Tas,
The EDataType stereotype extends DataType as of UML2 2.1.
Kenn
"Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in message
news:f0qn0t$bme$1@build.eclipse.org...
> Hi Kenn,
>
> It would help becuase Instance Class Name is what I am after, but on the
> profile I'm using I can't apply <<eClass>> to a DataType. I am using UML
> 2.0.2.. has this changed since then?
>
> Tas
>
>
> "Kenn Hussey" <khussey@ca.ibm.com> wrote in message
> news:f0qmcp$ace$1@build.eclipse.org...
>> Tas,
>>
>> Seeing as only PrimitiveType from UML maps to EDataType from Ecore, I
>> think it makes sense for the EDataType stereotype extend only
>> PrimitiveType. Note that you can stereotype a UML data type as <<eClass>>
>> if you want to specify some of the tags that are available from the
>> EDataType stereotype...
>>
>> Kenn
>>
>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>> message news:f0qfvo$ug9$1@build.eclipse.org...
>>> Hi James,
>>>
>>> I believe I understand the thinking there and it may be one of those
>>> highly subjective areas; What is considered primtive may depend on
>>> what's being modeled.
>>>
>>> Some other itneresting poitns are:
>>> * A PrimitiveType can have properties and operations
>>> * A Date (and many other standard java classes) have substructure yet
>>> you need to model them as primitves to use UML2Ecore.
>>>
>>> Do you think it would be contentious to allow the EDataType Stereotype
>>> to be applicable to both DataType and PrimitiveType?
>>>
>>> Thanks,
>>> Tas
>>>
>>>
>>> "James Bruck" <jbruck@ca.ibm.com> wrote in message
>>> news:f0qcvr$567$2@build.eclipse.org...
>>>> Hi Tas,
>>>>
>>>> That's a good question ... there definitely was some back and forth
>>>> when
>>>> this decision was made so it was not a clear cut decision.
>>>> At one point the XMLPrimitiveTypes library used UML DataTypes as
>>>> opposed to
>>>> Primitive types.
>>>>
>>>> The key point is that is that PrimitiveTypes are DataTypes but ones
>>>> that
>>>> don't have substructure ( see section 7.3.43 ).
>>>>
>>>> This maps more closely to EDataTypes in ecore.
>>>>
>>>> Regards,
>>>>
>>>> - James.
>>>>
>>>>
>>>> "Tas Frangoullides" <tas.frangoullides@barclaysglobal.com> wrote in
>>>> message
>>>> news:f0pr0n$las$1@build.eclipse.org...
>>>>> Hi all,
>>>>>
>>>>> Is there a partciular reason that EDataType cannot be applied to a UML
>>>>> DataType. I often use DataType in my models becuase it feels like the
>>>> right
>>>>> choice at that level of abstraction. If I need to process models
>>>>> through
>>>>> UML2Ecore I use my own profile and stereotype and trasnform the model
>>>> first,
>>>>> replacing DataType with Primitive Type + EDataType stereotype.. This
>>>>> is
>>>> fine
>>>>> but it would *nice* if I didn't have to do it.
>>>>>
>>>>> Tas.
>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>
>
|
|
|
Goto Forum:
Current Time: Fri Apr 26 22:28:36 GMT 2024
Powered by FUDForum. Page generated in 0.05019 seconds
|