Eclipse Community Forums: UML2 » Should the OCL allInstances() feature apply to stereotypes?

Help

Home

Home » Modeling » UML2 » Should the OCL allInstances() feature apply to stereotypes?

Show: Today's Messages :: Show Polls :: Message Navigator

Should the OCL allInstances() feature apply to stereotypes? [message #477610]

Wed, 16 July 2008 18:51

Linda Damus

Messages: 85
Registered: July 2009

Member

Hi,

Should the OCL allInstances() feature apply to stereotypes?

In other words, if I've created a stereotype 'S' extending Class in
profile 'P', should I be able to write an OCL constraint that uses the
expression:

P::S.allInstances()

and expect it to find all of the Classes in my model to which 'S' has
been applied?

Although the UML v2.1.2 specification talks about instances of an
instance "S" of Stereotype (section 18.3.8, semantics), it's not clear
to me whether or not a Class stereotyped as 'S' should or should not be
considered an instance of 'S'.

Thanks,
Linda

Report message to a moderator

Re: Should the OCL allInstances() feature apply to stereotypes? [message #477611 is a reply to message #477610]

Wed, 16 July 2008 19:18

Eclipse User

Originally posted by: cdamus.zeligsoft.com

--=-B9rx7rMixmoSavi/KItu
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi, Linda,

>From an OCL (and UML) perspective, Stereotypes are just Classes. The
metaclass extension is an association, and in OCL, is navigated just
like any other association.

The instances of stereotypes are elements associated with the instances
of the extended metaclasses. UML uses a triangular arrow-head for the
metaclass extension to indicate that the intention is to extend the
attributes of a UML element in a way like generalization (which also
uses a triangle), but I think it's clear that extensions do not actually
have generalization semantics. For example, the
Classifier::conformsTo(Type) operation does not consider a metaclass to
conform to an applied stereotype because there is no generalization.

So, allInstances() on a stereotype should only find the stereotype
"applications" and not the UML elements that they extend. However, from
the applications, it's easy enough to navigate to the UML elements via
the base_Xyz extension end property. Try this:

P::S.allInstances().base_Actor

as an example of getting a collection of Actors that have S applied if S
extends the Actor metaclass.

I don't know that this actually works in the MDT OCL implementation ...

HTH,

Christian

On Wed, 2008-07-16 at 14:51 -0400, Linda Damus wrote:

> Hi,
>
> Should the OCL allInstances() feature apply to stereotypes?
>
> In other words, if I've created a stereotype 'S' extending Class in
> profile 'P', should I be able to write an OCL constraint that uses the
> expression:
>
> P::S.allInstances()
>
> and expect it to find all of the Classes in my model to which 'S' has
> been applied?
>
> Although the UML v2.1.2 specification talks about instances of an
> instance "S" of Stereotype (section 18.3.8, semantics), it's not clear
> to me whether or not a Class stereotyped as 'S' should or should not be
> considered an instance of 'S'.
>
> Thanks,
> Linda

--=-B9rx7rMixmoSavi/KItu
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.16.0">
</HEAD>
<BODY>
Hi, Linda, 
 
>From an OCL (and UML) perspective, Stereotypes are just Classes.  The metaclass extension is an association, and in OCL, is navigated just like any other association. 
 
The instances of stereotypes are elements associated with the instances of the extended metaclasses.  UML uses a triangular arrow-head for the metaclass extension to indicate that the intention is to extend the attributes of a UML element in a way like generalization (which also uses a triangle), but I think it's clear that extensions do not actually have generalization semantics.  For example, the Classifier::conformsTo(Type) operation does not consider a metaclass to conform to an applied stereotype because there is no generalization. 
 
So, allInstances() on a stereotype should only find the stereotype "applications" and not the UML elements that they extend.  However, from the applications, it's easy enough to navigate to the UML elements via the base_Xyz extension end property.  Try this: 
 
  P::S.allInstances().base_Actor 
 
as an example of getting a collection of Actors that have S applied if S extends the Actor metaclass. 
 
I don't know that this actually works in the MDT OCL implementation ... 
 
HTH, 
 
Christian 
 
 
On Wed, 2008-07-16 at 14:51 -0400, Linda Damus wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
Hi,

Should the OCL allInstances() feature apply to stereotypes?

In other words, if I've created a stereotype 'S' extending Class in 
profile 'P', should I be able to write an OCL constraint that uses the 
expression:

P::S.allInstances()

and expect it to find all of the Classes in my model to which 'S' has 
been applied?

Although the UML v2.1.2 specification talks about instances of an 
instance "S" of Stereotype (section 18.3.8, semantics), it's not clear 
to me whether or not a Class stereotyped as 'S' should or should not be 
considered an instance of 'S'.

Thanks,
Linda
</PRE>
</BLOCKQUOTE>
</BODY>
</HTML>

--=-B9rx7rMixmoSavi/KItu--

Report message to a moderator

Re: Should the OCL allInstances() feature apply to stereotypes? [message #477612 is a reply to message #477611]

Wed, 16 July 2008 20:09

Linda Damus

Messages: 85
Registered: July 2009

Member

Thanks for the prompt answer, Christian!

I've tried your suggested:

P::S.allInstances().base_Actor

but it is always empty, even when 'S' is applied to Actors in my model.

I've opened bugzilla
https://bugs.eclipse.org/bugs/show_bug.cgi?id=241148 to request that
allInstances() return instances of stereotype "applications."

In the mean time, a combination of Actor.allInstances() and
getAppliedStereotype('P::S') will do the trick to express the constraint.

Thanks,
Linda

Christian W. Damus wrote:
> Hi, Linda,
>
> >From an OCL (and UML) perspective, Stereotypes are just Classes. The
> metaclass extension is an association, and in OCL, is navigated just
> like any other association.
>
> The instances of stereotypes are elements associated with the instances
> of the extended metaclasses. UML uses a triangular arrow-head for the
> metaclass extension to indicate that the intention is to extend the
> attributes of a UML element in a way like generalization (which also
> uses a triangle), but I think it's clear that extensions do not actually
> have generalization semantics. For example, the
> Classifier::conformsTo(Type) operation does not consider a metaclass to
> conform to an applied stereotype because there is no generalization.
>
> So, allInstances() on a stereotype should only find the stereotype
> "applications" and not the UML elements that they extend. However, from
> the applications, it's easy enough to navigate to the UML elements via
> the base_Xyz extension end property. Try this:
>
> P::S.allInstances().base_Actor
>
> as an example of getting a collection of Actors that have S applied if S
> extends the Actor metaclass.
>
> I don't know that this actually works in the MDT OCL implementation ...
>
> HTH,
>
> Christian
>
>
> On Wed, 2008-07-16 at 14:51 -0400, Linda Damus wrote:
>> Hi,
>>
>> Should the OCL allInstances() feature apply to stereotypes?
>>
>> In other words, if I've created a stereotype 'S' extending Class in
>> profile 'P', should I be able to write an OCL constraint that uses the
>> expression:
>>
>> P::S.allInstances()
>>
>> and expect it to find all of the Classes in my model to which 'S' has
>> been applied?
>>
>> Although the UML v2.1.2 specification talks about instances of an
>> instance "S" of Stereotype (section 18.3.8, semantics), it's not clear
>> to me whether or not a Class stereotyped as 'S' should or should not be
>> considered an instance of 'S'.
>>
>> Thanks,
>> Linda

Report message to a moderator

Re: Should the OCL allInstances() feature apply to stereotypes? [message #477613 is a reply to message #477612]

Wed, 16 July 2008 20:32

Eclipse User

Originally posted by: cdamus.zeligsoft.com

--=-S0idtfB5W7lMzty7yaD0
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi, Linda,

You should also be able to do this:

Actor.allInstances()->select(extension_S->notEmpty())

to get all of the Actors that have S applied, in an OCL-standard way
(the getAppliedStereotype(String) operation is not UML-compliant).

Or, if you need the S instances:

Actor.allInstances().extension_S

cW

On Wed, 2008-07-16 at 16:09 -0400, Linda Damus wrote:

> Thanks for the prompt answer, Christian!
>
> I've tried your suggested:
>
> P::S.allInstances().base_Actor
>
> but it is always empty, even when 'S' is applied to Actors in my model.
>
> I've opened bugzilla
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=241148 to request that
> allInstances() return instances of stereotype "applications."
>
> In the mean time, a combination of Actor.allInstances() and
> getAppliedStereotype('P::S') will do the trick to express the constraint.
>
> Thanks,
> Linda
>
>
> Christian W. Damus wrote:
> > Hi, Linda,
> >
> > >From an OCL (and UML) perspective, Stereotypes are just Classes. The
> > metaclass extension is an association, and in OCL, is navigated just
> > like any other association.
> >
> > The instances of stereotypes are elements associated with the instances
> > of the extended metaclasses. UML uses a triangular arrow-head for the
> > metaclass extension to indicate that the intention is to extend the
> > attributes of a UML element in a way like generalization (which also
> > uses a triangle), but I think it's clear that extensions do not actually
> > have generalization semantics. For example, the
> > Classifier::conformsTo(Type) operation does not consider a metaclass to
> > conform to an applied stereotype because there is no generalization.
> >
> > So, allInstances() on a stereotype should only find the stereotype
> > "applications" and not the UML elements that they extend. However, from
> > the applications, it's easy enough to navigate to the UML elements via
> > the base_Xyz extension end property. Try this:
> >
> > P::S.allInstances().base_Actor
> >
> > as an example of getting a collection of Actors that have S applied if S
> > extends the Actor metaclass.
> >
> > I don't know that this actually works in the MDT OCL implementation ...
> >
> > HTH,
> >
> > Christian
> >
> >
> > On Wed, 2008-07-16 at 14:51 -0400, Linda Damus wrote:
> >> Hi,
> >>
> >> Should the OCL allInstances() feature apply to stereotypes?
> >>
> >> In other words, if I've created a stereotype 'S' extending Class in
> >> profile 'P', should I be able to write an OCL constraint that uses the
> >> expression:
> >>
> >> P::S.allInstances()
> >>
> >> and expect it to find all of the Classes in my model to which 'S' has
> >> been applied?
> >>
> >> Although the UML v2.1.2 specification talks about instances of an
> >> instance "S" of Stereotype (section 18.3.8, semantics), it's not clear
> >> to me whether or not a Class stereotyped as 'S' should or should not be
> >> considered an instance of 'S'.
> >>
> >> Thanks,
> >> Linda

--=-S0idtfB5W7lMzty7yaD0
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.16.0">
</HEAD>
<BODY>
Hi, Linda, 
 
You should also be able to do this: 
 
  Actor.allInstances()->select(extension_S->notE mpty()) 
 
to get all of the Actors that have S applied, in an OCL-standard way (the getAppliedStereotype(String) operation is not UML-compliant). 
 
Or, if you need the S instances: 
 
  Actor.allInstances().extension_S 
 
cW 
 
On Wed, 2008-07-16 at 16:09 -0400, Linda Damus wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
Thanks for the prompt answer, Christian!

I've tried your suggested:

 P::S.allInstances().base_Actor

but it is always empty, even when 'S' is applied to Actors in my model.

I've opened bugzilla 
<A HREF="https://bugs.eclipse.org/bugs/show_bug.cgi?id=241148">https://bugs.eclipse.org/bugs/show_bug.cgi?id=241148</A> to request that 
allInstances() return instances of stereotype "applications."

In the mean time, a combination of Actor.allInstances() and 
getAppliedStereotype('P::S') will do the trick to express the constraint.

Thanks,
Linda

Christian W. Damus wrote:
> Hi, Linda,
> 
> >From an OCL (and UML) perspective, Stereotypes are just Classes. The 
> metaclass extension is an association, and in OCL, is navigated just 
> like any other association.
> 
> The instances of stereotypes are elements associated with the instances 
> of the extended metaclasses. UML uses a triangular arrow-head for the 
> metaclass extension to indicate that the intention is to extend the 
> attributes of a UML element in a way like generalization (which also 
> uses a triangle), but I think it's clear that extensions do not actually 
> have generalization semantics. For example, the 
> Classifier::conformsTo(Type) operation does not consider a metaclass to 
> conform to an applied stereotype because there is no generalization.
> 
> So, allInstances() on a stereotype should only find the stereotype 
> "applications" and not the UML elements that they extend. However, from 
> the applications, it's easy enough to navigate to the UML elements via 
> the base_Xyz extension end property. Try this:
> 
> P::S.allInstances().base_Actor
> 
> as an example of getting a collection of Actors that have S applied if S 
> extends the Actor metaclass.
> 
> I don't know that this actually works in the MDT OCL implementation ...
> 
> HTH,
> 
> Christian
> 
> 
> On Wed, 2008-07-16 at 14:51 -0400, Linda Damus wrote:
>> Hi,
>>
>> Should the OCL allInstances() feature apply to stereotypes?
>>
>> In other words, if I've created a stereotype 'S' extending Class in 
>> profile 'P', should I be able to write an OCL constraint that uses the 
>> expression:
>>
>> P::S.allInstances()
>>
>> and expect it to find all of the Classes in my model to which 'S' has 
>> been applied?
>>
>> Although the UML v2.1.2 specification talks about instances of an 
>> instance "S" of Stereotype (section 18.3.8, semantics), it's not clear 
>> to me whether or not a Class stereotyped as 'S' should or should not be 
>> considered an instance of 'S'.
>>
>> Thanks,
>> Linda
</PRE>
</BLOCKQUOTE>
</BODY>
</HTML>

--=-S0idtfB5W7lMzty7yaD0--

Report message to a moderator