RE: Invariants evaluating to undefined

RE: Invariants evaluating to undefined

From: Waldin, Earl ^lt;>
Date: Wed 15 Mar 2006 - 12:37:13 GMT
Message-ID: <>
The current OCL specification ptc/05-06-06 has a companion document
ptc/05-06-05 "FTF Report of the OCL 2.0 Finalization Task Force", which
describes all issues that were resolved in the current version of the
specficiation. When you see something in red like "Issue 5972:
InvalidType added" in the OCL spec, there is a corresponding issue in
the FTF report with that number. So you can look up at least a
discussion of what was done and, usually, why.
Earl Waldin

-----Original Message-----
[] On Behalf Of D H. Akehurst
Sent: Monday, 13 March, 2006 09:13
Subject: Re: Invariants evaluating to undefined

The  'Set{undefined} = Set{} ' issue is not a "proposal".
It was our interpretation of the information given in the standard. It
was the only interpretation we found to be consistent with 
"prop->isEmpty() = prop.oclIsUndefined()"
I have just looked at the June 2005 version of the standard,
in this there is explicit discussion of 'null' values, and their
containment in collections [section 8.2].
I have to say that I don't know if this is the latest version of the
standard, just the latest I could find on OMG web site.
I also don't really understand the new introduction of 'OclInvalid' Why
is this needed/different to OclUndefined ??


From: Dan CHIOREAN [] 
Sent: 11 March 2006 09:26
To:; D H. Akehurst;;; france@CS.ColoState.EDU;;;
Subject: Re: Invariants evaluating to undefined

Dear all,

OCL was conceived as a complementar formalism for the visual one, used
in modeling languages.  Therefore, even in cases of the most simple
models, like the model proposed by Jordi, representing the class diagram
is more that a wise attitude, it is mandatory.

OCL specifications must be as comprehensible and unambigous as possible.
In Jordi's example, the apropriate invariant is:

context Company
   inv maxSalaryDept:
      (not self.manager.oclIsUndefined) implies (self.maxSalaryDept >

This specification is more suggestive that self.manager->isEmpty.  Using
the oclIsUndefined operation, it is very clear that a department may
have at most one manager not a collection of managers.

In models, collections are obtained by navigating multiple associations.
The "->" operator can be used exclusively on collections.  My oppinion
is that using the ".property" expression instead of
"->collect(property)" was not the best decision in the OCL language.

Doing "automatic conversions" from an element to a collection of
elements is unacceptable; Element <> Collection(Element) - OCL is a
typed language.  Dave's proposal to consider Set{undefined} = Set{} may
lead to absurd situations.

Let consider the following case:  A Department may have at most one
manager and two spokesmen.  Both the manager and the spokesmen are
persons and the class Person has an attribute name.  Supose that at a
moment, the Department has a manager and has not spokesmen.  The
manager's name is unknown at (for) that moment.

Conforming to Dave's proposal, the expresion defined in the Department

context Department =, would be evaluated to
Set{undefined} = Set{}, resulting true. :((

In the following I would like to discuss other aspects related to
Jordi's question.

In the DepartmentPerson model, maxSalaryDept, is not an apropriate
invariant for the Department context because it depends on salary, an
attribute specified in the class Person.

In the wellknown CompanyExample model (used in all OCL specifications),
salary is natural modeled as an attribute of the associationClass Job.
My oppinion is that in Jordi's problem, the natural solution is to
specify a precondition for the addNewJob(p:Person, j:Job):Set(Job)
operation, specified in the Department class.  In case of an
updateSalary(ns:Integer) operation, specified in the class Job, a
similar precondition will be required.

Please find attached an archive containing some models and different OCL
specifications realized using OCLE.  The specifications may be evaluated
on different snapshots.

Best regards,

Dan Chiorean 


Yahoo! Mail
Bring photos to life! New
<**http://photomail> PhotoMail makes sharing a breeze. 
Received on Wed Mar 15 12:37:22 2006