3C

1
3C
3C
3C
Clear
Clean
Concise
2
UML 2, exactly

• Financial Systems Architects
• Mercury Computer SystemsMEGA InternationalTogeth
  erSoft
• HitachiOpenIT
• Cap Gemini Ernst YoungProf. Alain Wegmann

3
NOTICE
4
NOTICE

• Please interrupt with questions.

5
NOTICE

• Please interrupt with questions.
• Please.

6
Progress on UML 2

• The 3C submitters thank you for your feedback
  during our open process.

7
Progress on UML 2

• The 3C submitters thank you for your feedback
  during our open process.
• Weve made improvements and corrections. There
  is more to do.

8
Progress on UML 2

• The 3C submitters thank you for your feedback
  during our open process.
• Weve made improvements and corrections. There
  is more to do.
• The 3C submitters thank the U2P submitters for
  posting versions of their submission, as their
  work progresses.

9
Progress on UML 2

• The 3C submitters thank the U2P submitters for
  the invitation to consider working with them.

10
Progress on UML 2

• The 3C submitters thank the U2P submitters for
  the invitation to consider working with them.
• We replied Lets try it and see how it goes.

11
Progress on UML 2

• The 3C submitters thank the U2P submitters for
  the invitation to consider working with them.
• We replied Lets try it and see how it goes.
• Weve continued the dialog with a document.

12
Progress on several fronts

• Redefinable, Increment
• RoleBinding
• Marking, Geometry
• Association, Role, Type
• Metamodel drawings
• MOF

13
Progress on several fronts

• Redefinable
• Increment
• 3C will incorporate these valuable concepts.

14
Progress on several fronts

• RoleBinding
• This is one of several 3C conceptsnot included
  in the 3C submission as metamodel elements.

15
Progress on several fronts

• RoleBinding
• This is one of several 3C conceptsnot included
  in the 3C submission as metamodel elements.
• We understand this is essential for tools.We
  will add this under model management.

16
Progress on several fronts

• Marking
• We have added the concept, MDA marking, to model
  management.

17
Progress on several fronts

• Geometry
• 3C will add this concept.
• We see it as a placeholder for the diagram
  interchange data.

18
Progress on several fronts

• Association, Role, Type
• There has been convergence on these concepts.

19
Progress on several fronts

• MOF
• It seems that there is agreement of all
  submitters that the MOF kernel must be identical
  with the UML kernel.

20
Progress on several fronts

• MOF
• It seems that there is agreement of all
  submitters that the MOF kernel must be identical
  with the UML kernel.
• Other submitter may say this differently

21
Progress on several fronts

• MOF
• It seems that there is agreement of all
  submitters that the MOF kernel must be identical
  with the UML kernel.
• Other submitter may say this differentlyif
  their statement is stronger

22
Progress on several fronts

• MOF
• It seems that there is agreement of all
  submitters that the MOF kernel must be identical
  with the UML kernel.
• Other submitter may say this differentlyif
  their statement is stronger, we agree.

23
Progress on Association

• If there is no retreat from version 651,
  Association will be first class in U2P.

24
Progress on Association

• If there is no retreat from version 651,
  Association will be first class in U2P.
• August 2001 Association is a GeneralizableElement
  (but not a Classifier)

25
Progress on Association

• If there is no retreat from version 651,
  Association will be first class in U2P.
• August 2000 Association is a GeneralizableElement
  (but not a Classifier)
• December 2001 Association is a Classifier.

26
Clean

• We hope all submitters can agree tocome clean
  all the way on this.

27
Clean

• We hope all submitters can agree tocome clean
  all the way on this.
• In the 3C submission, an association has roles,
  and is not required to be bound to a particular
  class.

28
Progress on Role

• August 2001 Part is an AssociationEnd.
• October 2001 Part represents a set of instances
  owned by a Classifier.
• December 2001 Part represents a set of instances
  owned by an instance of a Classifer.

29
Clear

• We were slow in grasping the 651 concept, Part.

30
Clear

• Thanks to a clear, clean, and concise explanation
  by Bran Selic,we now understand.

31
Clear

• In 651, a part is a variable, to which different
  instances can be assigned.

32
Clear

• In 651, a part is a variable, to which different
  instances can be assigned.
• That variable is owned by a Classifier.

33
Clear

• In 651, a part is a variable, to which different
  instances can be assigned.
• That variable is owned by a Classifier.
• Hence, perhaps, the name, part.

34
Clear

• In 651, a part is a variable, to which different
  instances can be assigned.
• That variable is owned by a Classifier.
• Hence, perhaps, the name, part.
• This is the concept 3C calls role.

35
Clear

• In 651, a part is a variable, owned by a
  Classifierto which instances can be assigned.

36
Clear

• In 651, a part is a variable, owned by a
  Classifierto which instances can be assigned.
• 3C uses part for a role

37
Clear

• In 651, a part is a variable, owned by a
  Classifierto which instances can be assigned.
• 3C uses part for a rolein any part-whole
  correspondence.

38
Clear

• In 651, a part is a variable, owned by a
  Classifierto which instances can be assigned.
• 3C uses part for a rolein any part-whole
  correspondence.
• 3C uses role for a variable

39
Clear

• In 651, a part is a variable, owned by a
  Classifierto which instances can be assigned.
• 3C uses part for a rolein any part-whole
  correspondence.
• 3C role for a variablewherever used in the
  model.

40
Progress on Type

• It is becoming clear that the proposal to make
  type a first class concept is being accepted in
  the U2P submission.

41
Progress on Type

• It is becoming clear that the proposal to make
  type a first class concept is being accepted in
  the U2P submission.
• Many elements of 651 have types.

42
Progress on Type

• The proposal to make type a first class concept
  is creeping in to the U2P submission.
• Many elements of 651 have types.
• These types are all Classifiers.

43
Progress on Type

• 3C type may be very close to one of the meanings
  of 651 Classifier.concept is creeping in to the
  U2P submission.
• Many elements of 651 have types.
• These types are all Classifiers.

44
Progress on Type

• 3C type may be very close to one of the meanings
  of 651 Classifier.concept is creeping in to the
  U2P submission.
• Differences remain3C UML has types for more
  than the concepts that fall under 651
  Classifier.

45
Concise

• 3C has a single term for each kindof modeling
  concept

46
Concise

• 3C has a single term for each kindof modeling
  concept
• 3C has,

47
Concise

• 3C has a single term for each kindof modeling
  concept
• 3C has, once,

48
Concise

• 3C has a single term for each kindof modeling
  concept
• 3C has, once, for all kinds of concepts,

49
Concise

• 3C has a single term for each kindof modeling
  concept
• 3C has, once, for all kinds of concepts,type

50
Concise

• 3C has a single term for each kindof modeling
  concept
• 3C has, once, for all kinds of concepts,typeand
  implementation.

51
Concise

• Reminder
• A classdeclares an object typeand specifies its
  implementation
• As, for example, in Java.

52
Concise

• Reminder
• A classdeclares an object typeand specifies its
  implementation
• No problem

53
Concise

• Reminder
• A classdeclares an object typeand specifies its
  implementation
• No problem Class is a concept that isa
  combination of two other concepts

54
Concise

• Reminder
• A classdeclares an object typeand specifies its
  implementation
• No problem Class is a concept that isa
  combination of two other concepts type
  and implementation

55
Lack of progress

• Namespace, Package, Permission
• Metamodel drawings
• OCL

56
Lack of progress

• Namespace, Package, Permission
• We have not done our work in this area.

57
Lack of progress

• Namespace, Package, Permission
• We have not done our work in this area.
• (Of course, 3C does include Namespace and
  Package. These are consistent with other
  submissions, but lack their detail.)

58
Lack of progress

• Namespace, Package, Permission
• We have not done our work in this area.
• We expect every effort will be made to see that
  these are identical in MOF and UML

59
Lack of progress

• OCL
• We have OCL in our infrastructure submission,
  but not a lot.

60
Lack of progress

• OCL
• We have OCL in our infrastructure submission,
  but not a lot.
• 2U, of course, is generating most of their OCL
  automatically.

61
Lack of progress

• OCL
• We have OCL in our infrastructure submission,
  but not a lot.
• 2U, of course, is generating most of their OCL
  automatically.
• U2P does not have OCL yet.

62
Lack of progress

• Metamodel drawings
• We have made progress on drawings.
• Next week we start the work of checking them for
  MOF compliance.

63
Some other highlights
64
Message as role

• In UML 1 and in U2P 651, Message is an added
  concept ( Model Element in 1, NamedElement in
  651).

We model a message as a role in an action.
In other kinds of CORBA operation invocations
four roles are client, object, provided message
andreturned message.
For example, in a CORBA one-way operation
invocation three roles are client, object and
provided message.
65
Action types

• Types of joint action, not decomposed, specified
  using roles

66
Action types

• Types of joint action, not decomposed, specified
  using roles
• A multimedia flow

67
Action types

• Types of joint action, not decomposed, specified
  using roles
• A multimedia flow
• A set of distributed systems keep their clocks
  nearly synchronized

68
Action types

• Types of joint action, not decomposed, specified
  using roles
• A multimedia flow
• A set of distributed systems keep their clocks
  nearly synchronized
• Three objects, two peers and a certification
  authority, establish an authenticated and secure
  association between the two peers

69
A couple of the meanings

• Note that a user model contains instance-of
  relationships between Classes and their
  objects.

The instance-of relationship is used
within a (meta)model,
while the instance-of relationship is used
across metalayers. U2P 651
70
One word - Five meanings

• the language use relation
• 4 This model is an instance of the UML
  metamodel.
• 1 This model uses UML. I like my house.
  is English

71
One word - Five meanings

• the meta-linguistic relation
• 4 Customer is an instance of Class.
• 1 Customer is a class house is a noun.

72
One word - Five meanings

• the instantitation relation
• 4 aCustomer is an instance of Customer.
• 1 aCustomer was instantiated from the
  class, Customer. My house was built from these
  plans.

73
One word - Five meanings

• the type conformance relation
• 4 aCustomer is an instance of Customer.
• 1 aCustomer is an instance of the object
  type, NewCustomer. This room is a closet.

74
One word - Five meanings

• the representation relation
• 4 aCustomer is an instance of Customer.
• 1 aCustomer represents a typical customer.
  This symbol represents
  that door there.

75
Object

• 3C is backwards, starting with object.

76
Object

• 3C is backwards, starting with object.
• (So is OP, the SightCode submission)

77
Object

• 3C is backwards, starting with object.
• (So is OP, the SightCode submission)
• 2U starts with class and uses object in their
  semantic domain.

78
Object

• 3C is backwards, starting with object.
• (So is OP, the SightCode submission)
• 2U starts with class and uses object in their
  semantic domain.
• U2P 651 does not have the concept, Object.

79
Recognized?

• We recognize that there are theoretical limits
  to what one can express about a metamodel using
  the meta-model itself.

80
Family of Languages

• We are convinced that the family of languages
  idea is central.
• So we have placed the extension mechanism in the
  kernel.

81
Supporters

• Responsibilities feedback

82
Supporters

• Responsibilities feedback
• Access (sorry) no improvement in access

83
Supporters

• Responsibilities feedback
• Access (sorry) no improvement in access
• We feel there is no commercial reasonto develop
  UML in secret.

84
Collaboration

• We have not received any requests for specific
  changes to our submission

85
Collaboration

• We have not received any requests for specific
  changes to our submission from other submission
  groups.

86
Collaboration

• We have not received any requests for specific
  changes to our submission from other submission
  groups.
• Of course, we welcome requests.

87
UML 2, exactly

• Financial Systems Architects
• Mercury Computer SystemsMEGA InternationalTogeth
  erSoft
• HitachiOpenIT
• Cap Gemini Ernst YoungProf. Alain Wegmann
• Responsibilities feedback
• Bandwidth (sorry) the same.

88

• 3C
• www.community-ML.org

89
Thats all, folks!