Issue 4407: Inconsistencies caused by changing Expression etc from Data Types to Classe (cwm-rtf)
Source: International Business Machines (Dr. Daniel T. Chang, dtchang(at)us.ibm.com)
Nature: Revision
Severity: Significant
Summary: Expression and their subtypes (BooleanExpression, etc.) were changed from Data Types (in the CWM Adapted Specification) to Classes in CWM 1.0. As a result, it caused design inconsistency in CWM. For example, ExpressionNode inherits from Element. This was designed originally based on the fact that Expression was a Data Type and could not be subclassed. It should now inherit from Expression, which can be subclassed. The CWM RTF should review and revise all such inconsistencies caused by changing Expression, Multiplicity, etc. from Data Types to Classes. 

Resolution: 
Revised Text: 
Actions taken:
July 24, 2001: received issue
Discussion: 

End of Annotations:=====
from: webmaster@omg.org
Message-Id: <200107241759.f6OHxtt12018@emerald.omg.org>
Date: 24 Jul 2001 14:01:34 -0400
To: <juergen@omg.org>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Subject: Issue/Bug Report
Content-Type: Text/html; charset=windows-1252
X-UIDL: FU)e9&7nd9"Dp!!aaM!!



Name: Dan Chang 
Company: IBM 
mailFrom: dtchang@us.ibm.com 
Notification: No 
Specification: Common Warehouse Metamodel (CWM) Specification 
Section: 7.3.2 
FormalNumber: ad/01-02-01 
Version: 1.0 
RevisionDate: 2 Feb 2001 
Page: 7-70 
Nature: Revision 
Severity: Significant 
HTTP User Agent: Mozilla/4.61 [en] (WinNT; U) 


Description

Expression and their subtypes (BooleanExpression, etc.) were changed
from Data Types (in the CWM Adapted Specification) to Classes in CWM
1.0. As a result, it caused design inconsistency in CWM. For example,
ExpressionNode inherits from Element. This was designed originally
based on the fact that Expression was a Data Type and could not be
subclassed. It should now inherit from Expression, which can be
subclassed. The CWM RTF should review and revise all such
inconsistencies caused by changing Expression, Multiplicity, etc. from
Data Types to Classes. 

From: "Tolbert, Doug M" <Doug.Tolbert@unisys.com>
To: cwm-rtf@omg.org
Subject: "Baisley" questions from Danvers CWM RTF meeting
Date: Wed, 18 Jul 2001 18:18:03 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: 2mVd9@Me!!'mi!!*6b!!

I had a chance to discuss some of the questions that surface at last
Thursday's RTF meeting with Don Baisley.  Herewith, his responses:

(1) Interpretation of the meaning of the <<metamodel>> stereotype.  In
    our
review of the proposed Data Mining changes there was some confusion
    overview
the meaning of the <<metamodel>> stereotype and its effect on package
structure during generation.  The DM team was trying to use the
    stereotype
at multiple levels within their proposed package structure as a way of
achieving optionality of some subpackages.  In contrast, I had
    contended
that <<metamodel>> stereotypes could not be nested it this way.  Turns
    out
that I was (mostly) correct ... the <<metamodel>> stereotype should be
    used
at the lowest level of granularity where creation of a new package by
    the
generator is desired.  It is possible (but untested) that nested
<<metamodel>> stereotypes would not create error situations because
    the Rose
extractor is believed to simply ignore nested <<metamodel>>
    stereotypes.  In
short, the outermost <<metamodel>> stereotypes are used to defined the
package structure (ie, defines the boundry of an XML extent); any
    nested
<<metamodel>> stereotypes will (probably) be ignored.  Don did say
    that
nesting <<metamodel>> stereotypes was definitely NOT recommended
    usage.  So,
packages that are mutually optional must both have <<metamodel>>
stereotypes.  Note, however, the nesting packages themselves IS
    recommended
usage -- packages are preserved at all levels, <<metamodel>>
    stereotype or
not.

(2) Dan asked if CIM would support generation XML Schemas for CWM in
addition to XML DTDs.  The strict answer is "no." However, Don
recommended
that Dan should consult with Steve Brodsky about the available of XML
Schema
generation from the XMI team.

(3) Expressions as datatypes.  At first blush, Don doesn't think this
    is
illegal (as was discussed in Danvers).  Consequently, I've posted an
    issue
for the RTF on the OMG issues list about this topic so that it can be
considered more fully.  Don and I will take a more detailed look at it
    in
the near future.

Doug

From: "Iyengar, Sridhar" <Sridhar.Iyengar2@unisys.com>
To: Dan Chang <dtchang@us.ibm.com>,
   "Tolbert, Doug M"
    <Doug.Tolbert@unisys.com>
Cc: cwm-rtf@omg.org, "Iyengar, Sridhar" <Sridhar.Iyengar2@unisys.com>
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Date: Wed, 18 Jul 2001 20:49:37 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: Vh?e9nM>e9QGd!!^h4!!

Regarding Dan's question on Unisys help on IDL/XML generation - Here
is the
Unisys position.

Based on resources available we (Unisys) will assist in some of these
generations. But several vendors including IBM, Adaptive, SUN and
Oracle
have or will soon have this technology. So hopefully the xsd
generation is a
short term issue.  

P.S.  The ADTF has not yet voted on XMI2.  This will happen in
Toronto.

Sridhar



------------------------------------------------------------------------
----
Sridhar Iyengar
Unisys Fellow, Director of Advanced Technology
25725, Jeronimo Road
Mission Viejo, CA 92691

E-mail : Sridhar.iyengar2@unisys.com
Phone : 949-380-5692
Fax     : 949-380-6632
------------------------------------------------------------------------
-
> 
Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
To: "Tolbert, Doug M" <Doug.Tolbert@unisys.com>
Cc: cwm-rtf@omg.org, "Iyengar, Sridhar" <Sridhar.Iyengar2@unisys.com>
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OFECA99ADB.B4DD033B-ON88256A8D.00833F8F@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Wed, 18 Jul 2001 17:26:51 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
|June 18, 2001) at
 07/18/2001 06:26:52 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: Wf^!!\[(!!^^Ce9~$0e9


Doug,

My comment is embedded below (==>).

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                          
                    "Tolbert, Doug                                                                        
                    M"                    To:     cwm-rtf@omg.org                                         
                    <Doug.Tolbert@u       cc:                                                             
                    nisys.com>            Subject:     "Baisley"
                    questions from Danvers CWM RTF meeting   
                                                                                                          
                    07/18/01 04:18                                                                        
                    PM                                                                                    
                                                                                                          
                                                                                                          



I had a chance to discuss some of the questions that surface at last
Thursday's RTF meeting with Don Baisley.  Herewith, his responses:

(1) Interpretation of the meaning of the <<metamodel>> stereotype.  In
    our
review of the proposed Data Mining changes there was some confusion
overview
the meaning of the <<metamodel>> stereotype and its effect on package
structure during generation.  The DM team was trying to use the
    stereotype
at multiple levels within their proposed package structure as a way of
achieving optionality of some subpackages.  In contrast, I had
    contended
that <<metamodel>> stereotypes could not be nested it this way.  Turns
    out
that I was (mostly) correct ... the <<metamodel>> stereotype should be
    used
at the lowest level of granularity where creation of a new package by
    the
generator is desired.  It is possible (but untested) that nested
<<metamodel>> stereotypes would not create error situations because
    the
Rose
extractor is believed to simply ignore nested <<metamodel>>
    stereotypes.
In
short, the outermost <<metamodel>> stereotypes are used to defined the
package structure (ie, defines the boundry of an XML extent); any
    nested
<<metamodel>> stereotypes will (probably) be ignored.  Don did say
    that
nesting <<metamodel>> stereotypes was definitely NOT recommended
    usage.
So,
packages that are mutually optional must both have <<metamodel>>
stereotypes.  Note, however, the nesting packages themselves IS
    recommended
usage -- packages are preserved at all levels, <<metamodel>>
    stereotype or
not.
==> Agree. We should all use <<metamodel>> stereotype at the leaf
    level.

(2) Dan asked if CIM would support generation XML Schemas for CWM in
addition to XML DTDs.  The strict answer is "no." However, Don
recommended
that Dan should consult with Steve Brodsky about the available of XML
Schema
generation from the XMI team.
==> Does this mean that the Unisys subteam of CWM will no longer
provide
the valuable
==> service of generating XML/DTD/XSD/IDL files for future CWM
revisions,
starting with
==> CWM 1.1?

(3) Expressions as datatypes.  At first blush, Don doesn't think this
    is
illegal (as was discussed in Danvers).  Consequently, I've posted an
    issue
for the RTF on the OMG issues list about this topic so that it can be
considered more fully.  Don and I will take a more detailed look at it
    in
the near future.
==> Please see the last paragraph on page 56 of MOF 1.3: All
    (Attributes)
have a "type" that is
==> represented using a Data Type ..., and page 57 on "type:": This is
expressed as the name of
==> a Data Type defined in Section 3.6, or as the name of a CORBA data
    type
...
==> Therefore, it is illegal to use a CWM-defined Expression (an
    instance
of a MOF Class) as the
==> type in defining a CWM attribute (an instance of a MOF attribute).
==> I am copying Sridhar on this note so that he can confirm or rebuke
    what
are stated in MOF 1.3.

Doug

From: "Pete Rivett" <pete.rivett@adaptive.com>
To: "'Dan Chang'" <dtchang@us.ibm.com>,
   "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>
Cc: <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
<Sridhar.Iyengar2@unisys.com>
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Date: Thu, 19 Jul 2001 03:49:57 +0100
Message-ID: <002c01c10ffd$7f8faf30$114c04c8@CHIMAY>
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2911.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.3018.1300
In-Reply-To: <OFECA99ADB.B4DD033B-ON88256A8D.00833F8F@boulder.ibm.com>
Content-Type: text/plain;
	      charset="us-ascii"
X-UIDL: \`i!!@`Z!!gJ^!!QB7e9

Dan,
Re your comment on point 3):
The section you quote from MOF 1.3 is in the section about the MOF
Model
itself (the M3): it's in effect saying that although the MOF M3 is an
instance of itself, it is (voluntarily) restricting itself to only
defining
attributes based on Data Types.
This is placing no such constraint on M2's such as CWM.
The (M3) class 'Attribute' inherits the reference 'type' from
TypedElement
which refers to Classifier. For Attribute the only constraint on this
reference is at the TypedElement level where [C-13] states that "An
Association may not be the type of a TypedElement" (implicitly
allowing the
other 2 subtypes of Classifier - Class and DataType).

I'm not sure this addresses the original question since I was not
involved
in that discussion at Danvers: but CWM class 'Expression' may legally
be
used as the type of a CWM-defined attribute. Whether it's the best way
of
modeling the requirement is another question (e.g. it does not allow a
(possibly complex or volatile) Expression (e.g. "Sales tax in
California")
to be defined only once and referred to from many places, minimizing
the
impact of any change).

Pete

Pete Rivett (pete.rivett@adaptive.com)
Chief Technology Officer, Adaptive Ltd
Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL, UK
Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
http://www.adaptive.com



> -----Original Message-----
> From: Dan Chang [mailto:dtchang@us.ibm.com]
> Sent: 19 July 2001 01:27
> To: Tolbert, Doug M
> Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
> Doug,
>
> My comment is embedded below (==>).
>
> Regards,  Dan
>
> e-business Data Technology and Standard
> IBM Silicon Valley Laboratory
> Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> Internet:  dtchang@us.ibm.com
> VM:          IBMUSM50(DTCHANG)
> Phone:    (408)-463-2319
>
--cut--
>
>
> (3) Expressions as datatypes.  At first blush, Don doesn't
> think this is
> illegal (as was discussed in Danvers).  Consequently, I've
> posted an issue
> for the RTF on the OMG issues list about this topic so that it can
be
> considered more fully.  Don and I will take a more detailed
> look at it in
> the near future.
> ==> Please see the last paragraph on page 56 of MOF 1.3: All
> (Attributes)
> have a "type" that is
> ==> represented using a Data Type ..., and page 57 on "type:": This
is
> expressed as the name of
> ==> a Data Type defined in Section 3.6, or as the name of a
> CORBA data type
> ...
> ==> Therefore, it is illegal to use a CWM-defined Expression
> (an instance
> of a MOF Class) as the
> ==> type in defining a CWM attribute (an instance of a MOF
attribute).
> ==> I am copying Sridhar on this note so that he can confirm
> or rebuke what
> are stated in MOF 1.3.
>
> Doug
>
>
>
>
>
>
>


The information contained in this email and any attached files are
confidential and intended solely for the addressee(s). The e-mail may
be legally privileged or prohibited from disclosure and unauthorised
use.
If you are not the named addressee you may not use, copy or disclose
this information to any other person.  If you received this message in
error please notify the sender immediately.
Any views or opinions presented here may be solely those of the
originator and do not necessarily reflect those of the Company.


From: John_Poole@hyperion.com
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
To: "Pete Rivett" <pete.rivett@adaptive.com>
Cc: "'Dan Chang'" <dtchang@us.ibm.com>,
   "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>, <cwm-rtf@omg.org>,
   "'Iyengar, Sridhar'" <Sridhar.Iyengar2@unisys.com>
Date: Thu, 19 Jul 2001 10:47:41 -0400
Message-ID: <OFFA360E7C.F9784FE3-ON85256A8E.00503B59@hyperion.com>
X-MIMETrack: Serialize by Router on Stmfd-Gateway1/na/Hyperion(Release
5.0.5 |September
 22, 2000) at 07/19/2001 10:47:43 AM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: &QG!!e)e!!RAn!!IM'!!



Hi all,

It seems to me, by way of analogy, that if it is illegal for me to use
a
CWM Expression as an attribute type in another CWM class, then it
would be
equally illegal for me to use any other CWM class as the type of an
object-valued attribute in another CWM class (e.g., ExpressionNode).

Is this a reasonable conclusion, or am I mistaken about this?

- John






"Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001 10:49:57 PM

To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
      <Doug.Tolbert@unisys.com>
cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
      <Sridhar.Iyengar2@unisys.com>
Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting


Dan,
Re your comment on point 3):
The section you quote from MOF 1.3 is in the section about the MOF
Model
itself (the M3): it's in effect saying that although the MOF M3 is an
instance of itself, it is (voluntarily) restricting itself to only
defining
attributes based on Data Types.
This is placing no such constraint on M2's such as CWM.
The (M3) class 'Attribute' inherits the reference 'type' from
TypedElement
which refers to Classifier. For Attribute the only constraint on this
reference is at the TypedElement level where [C-13] states that "An
Association may not be the type of a TypedElement" (implicitly
allowing the
other 2 subtypes of Classifier - Class and DataType).

I'm not sure this addresses the original question since I was not
involved
in that discussion at Danvers: but CWM class 'Expression' may legally
be
used as the type of a CWM-defined attribute. Whether it's the best way
of
modeling the requirement is another question (e.g. it does not allow a
(possibly complex or volatile) Expression (e.g. "Sales tax in
California")
to be defined only once and referred to from many places, minimizing
the
impact of any change).

Pete

Pete Rivett (pete.rivett@adaptive.com)
Chief Technology Officer, Adaptive Ltd
Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL, UK
Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
http://www.adaptive.com



> -----Original Message-----
> From: Dan Chang [mailto:dtchang@us.ibm.com]
> Sent: 19 July 2001 01:27
> To: Tolbert, Doug M
> Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
> Doug,
>
> My comment is embedded below (==>).
>
> Regards,  Dan
>
> e-business Data Technology and Standard
> IBM Silicon Valley Laboratory
> Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> Internet:  dtchang@us.ibm.com
> VM:          IBMUSM50(DTCHANG)
> Phone:    (408)-463-2319
>
--cut--
>
>
> (3) Expressions as datatypes.  At first blush, Don doesn't
> think this is
> illegal (as was discussed in Danvers).  Consequently, I've
> posted an issue
> for the RTF on the OMG issues list about this topic so that it can
be
> considered more fully.  Don and I will take a more detailed
> look at it in
> the near future.
> ==> Please see the last paragraph on page 56 of MOF 1.3: All
> (Attributes)
> have a "type" that is
> ==> represented using a Data Type ..., and page 57 on "type:": This
is
> expressed as the name of
> ==> a Data Type defined in Section 3.6, or as the name of a
> CORBA data type
> ...
> ==> Therefore, it is illegal to use a CWM-defined Expression
> (an instance
> of a MOF Class) as the
> ==> type in defining a CWM attribute (an instance of a MOF
attribute).
> ==> I am copying Sridhar on this note so that he can confirm
> or rebuke what
> are stated in MOF 1.3.
>
> Doug
>
>
>
>
>
>
>


The information contained in this email and any attached files are
confidential and intended solely for the addressee(s). The e-mail may
be
legally privileged or prohibited from disclosure and unauthorised use.
If you are not the named addressee you may not use, copy or disclose
this
information to any other person.  If you received this message in
error
please notify the sender immediately.
Any views or opinions presented here may be solely those of the
originator
and do not necessarily reflect those of the Company.

From: "Pete Rivett" <pete.rivett@adaptive.com>
To: <John_Poole@hyperion.com>
Cc: "'Dan Chang'" <dtchang@us.ibm.com>,
   "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>, <cwm-rtf@omg.org>,
   "'Iyengar, Sridhar'" <Sridhar.Iyengar2@unisys.com>
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Date: Thu, 19 Jul 2001 16:20:05 +0100
Message-ID: <004201c11066$4a4096a0$114c04c8@CHIMAY>
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2911.0)
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.3018.1300
In-Reply-To: <OFFA360E7C.F9784FE3-ON85256A8E.00503B59@hyperion.com>
Content-Type: text/plain;
	      charset="us-ascii"
X-UIDL: 7bJ!!7PGe9TIO!!3=ed9

John,
I hope I made the point in my previous mail that it is NOT illegal to
use
CWM's Expression class as the type of an attribute.

Nonetheless your conditional statement is correct - although for
clarity I'd
replace "a CWM Expression" (which could be read as "an instance of the
class
Expression") with "CWM's Expression class".

Pete

> -----Original Message-----
> From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com]
> Sent: 19 July 2001 15:48
> To: Pete Rivett
> Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org;
> 'Iyengar, Sridhar'
> Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
>
> Hi all,
>
> It seems to me, by way of analogy, that if it is illegal for
> me to use a
> CWM Expression as an attribute type in another CWM class,
> then it would be
> equally illegal for me to use any other CWM class as the type of an
> object-valued attribute in another CWM class (e.g., ExpressionNode).
>
> Is this a reasonable conclusion, or am I mistaken about this?
>
> - John
>
>
>
>
>
>
> "Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001 10:49:57 PM
>
> To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
>       <Doug.Tolbert@unisys.com>
> cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
> <Sridhar.Iyengar2@unisys.com>
> Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
> Dan,
> Re your comment on point 3):
> The section you quote from MOF 1.3 is in the section about
> the MOF Model
> itself (the M3): it's in effect saying that although the MOF M3 is
an
> instance of itself, it is (voluntarily) restricting itself to
> only defining
> attributes based on Data Types.
> This is placing no such constraint on M2's such as CWM.
> The (M3) class 'Attribute' inherits the reference 'type' from
> TypedElement
> which refers to Classifier. For Attribute the only constraint on
this
> reference is at the TypedElement level where [C-13] states that "An
> Association may not be the type of a TypedElement"
> (implicitly allowing the
> other 2 subtypes of Classifier - Class and DataType).
>
> I'm not sure this addresses the original question since I was
> not involved
> in that discussion at Danvers: but CWM class 'Expression' may
> legally be
> used as the type of a CWM-defined attribute. Whether it's the
> best way of
> modeling the requirement is another question (e.g. it does not allow
a
> (possibly complex or volatile) Expression (e.g. "Sales tax in
> California")
> to be defined only once and referred to from many places,
> minimizing the
> impact of any change).
>
> Pete
>
> Pete Rivett (pete.rivett@adaptive.com)
> Chief Technology Officer, Adaptive Ltd
> Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL, UK
> Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
> http://www.adaptive.com
>
>
>
> > -----Original Message-----
> > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > Sent: 19 July 2001 01:27
> > To: Tolbert, Doug M
> > Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> > Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> >
> > Doug,
> >
> > My comment is embedded below (==>).
> >
> > Regards,  Dan
> >
> > e-business Data Technology and Standard
> > IBM Silicon Valley Laboratory
> > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > Internet:  dtchang@us.ibm.com
> > VM:          IBMUSM50(DTCHANG)
> > Phone:    (408)-463-2319
> >
> --cut--
> >
> >
> > (3) Expressions as datatypes.  At first blush, Don doesn't
> > think this is
> > illegal (as was discussed in Danvers).  Consequently, I've
> > posted an issue
> > for the RTF on the OMG issues list about this topic so that
> it can be
> > considered more fully.  Don and I will take a more detailed
> > look at it in
> > the near future.
> > ==> Please see the last paragraph on page 56 of MOF 1.3: All
> > (Attributes)
> > have a "type" that is
> > ==> represented using a Data Type ..., and page 57 on
> "type:": This is
> > expressed as the name of
> > ==> a Data Type defined in Section 3.6, or as the name of a
> > CORBA data type
> > ...
> > ==> Therefore, it is illegal to use a CWM-defined Expression
> > (an instance
> > of a MOF Class) as the
> > ==> type in defining a CWM attribute (an instance of a MOF
> attribute).
> > ==> I am copying Sridhar on this note so that he can confirm
> > or rebuke what
> > are stated in MOF 1.3.
> >
> > Doug
> >
> >
> >
> >
> >
> >
> >
>
>
> The information contained in this email and any attached files are
> confidential and intended solely for the addressee(s). The
> e-mail may be
> legally privileged or prohibited from disclosure and unauthorised
use.
> If you are not the named addressee you may not use, copy or
> disclose this
> information to any other person.  If you received this
> message in error
> please notify the sender immediately.
> Any views or opinions presented here may be solely those of
> the originator
> and do not necessarily reflect those of the Company.
>
>
>
>
>


The information contained in this email and any attached files are
confidential and intended solely for the addressee(s). The e-mail may
be legally privileged or prohibited from disclosure and unauthorised
use.
If you are not the named addressee you may not use, copy or disclose
this information to any other person.  If you received this message in
error please notify the sender immediately.
Any views or opinions presented here may be solely those of the
originator and do not necessarily reflect those of the Company.

From: John_Poole@hyperion.com
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
To: "Pete Rivett" <pete.rivett@adaptive.com>
Cc: "'Dan Chang'" <dtchang@us.ibm.com>,
   "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>, <cwm-rtf@omg.org>,
   "'Iyengar, Sridhar'" <Sridhar.Iyengar2@unisys.com>
Date: Thu, 19 Jul 2001 11:50:35 -0400
Message-ID: <OFFACCC3EC.97C93E1F-ON85256A8E.0056F517@hyperion.com>
X-MIMETrack: Serialize by Router on Stmfd-Gateway1/na/Hyperion(Release
5.0.5 |September
 22, 2000) at 07/19/2001 11:50:42 AM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: 8FAe9aA#!!QEi!!38Be9


Hi,

Yes, your point was clear.  I was merely concurring in a round-about
sort
of way......

- John

From: "Tolbert, Doug M" <Doug.Tolbert@unisys.com>
To: Pete Rivett <pete.rivett@adaptive.com>, John_Poole@hyperion.com
Cc: "'Dan Chang'" <dtchang@us.ibm.com>,
   "Tolbert, Doug M"
    <Doug.Tolbert@unisys.com>, cwm-rtf@omg.org,
   "Iyengar, Sridhar"
    <Sridhar.Iyengar2@unisys.com>
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Date: Thu, 19 Jul 2001 12:08:30 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: kijd99_3!!WkXd9f~&e9

Given the chaos #3 created, I think my choice of creating an issue was
the
right one!

-----Original Message-----
From: Pete Rivett [mailto:pete.rivett@adaptive.com]
Sent: Thursday, July 19, 2001 8:20 AM
To: John_Poole@hyperion.com
Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org; 'Iyengar,
Sridhar'
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting


John,
I hope I made the point in my previous mail that it is NOT illegal to
use
CWM's Expression class as the type of an attribute.

Nonetheless your conditional statement is correct - although for
clarity I'd
replace "a CWM Expression" (which could be read as "an instance of the
class
Expression") with "CWM's Expression class".

Pete

> -----Original Message-----
> From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com]
> Sent: 19 July 2001 15:48
> To: Pete Rivett
> Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org;
> 'Iyengar, Sridhar'
> Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
>
> Hi all,
>
> It seems to me, by way of analogy, that if it is illegal for
> me to use a
> CWM Expression as an attribute type in another CWM class,
> then it would be
> equally illegal for me to use any other CWM class as the type of an
> object-valued attribute in another CWM class (e.g., ExpressionNode).
>
> Is this a reasonable conclusion, or am I mistaken about this?
>
> - John
>
>
>
>
>
>
> "Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001 10:49:57 PM
>
> To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
>       <Doug.Tolbert@unisys.com>
> cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
> <Sridhar.Iyengar2@unisys.com>
> Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
> Dan,
> Re your comment on point 3):
> The section you quote from MOF 1.3 is in the section about
> the MOF Model
> itself (the M3): it's in effect saying that although the MOF M3 is
an
> instance of itself, it is (voluntarily) restricting itself to
> only defining
> attributes based on Data Types.
> This is placing no such constraint on M2's such as CWM.
> The (M3) class 'Attribute' inherits the reference 'type' from
> TypedElement
> which refers to Classifier. For Attribute the only constraint on
this
> reference is at the TypedElement level where [C-13] states that "An
> Association may not be the type of a TypedElement"
> (implicitly allowing the
> other 2 subtypes of Classifier - Class and DataType).
>
> I'm not sure this addresses the original question since I was
> not involved
> in that discussion at Danvers: but CWM class 'Expression' may
> legally be
> used as the type of a CWM-defined attribute. Whether it's the
> best way of
> modeling the requirement is another question (e.g. it does not allow
a
> (possibly complex or volatile) Expression (e.g. "Sales tax in
> California")
> to be defined only once and referred to from many places,
> minimizing the
> impact of any change).
>
> Pete
>
> Pete Rivett (pete.rivett@adaptive.com)
> Chief Technology Officer, Adaptive Ltd
> Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL, UK
> Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
> http://www.adaptive.com
>
>
>
> > -----Original Message-----
> > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > Sent: 19 July 2001 01:27
> > To: Tolbert, Doug M
> > Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> > Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> >
> > Doug,
> >
> > My comment is embedded below (==>).
> >
> > Regards,  Dan
> >
> > e-business Data Technology and Standard
> > IBM Silicon Valley Laboratory
> > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > Internet:  dtchang@us.ibm.com
> > VM:          IBMUSM50(DTCHANG)
> > Phone:    (408)-463-2319
> >
> --cut--
> >
> >
> > (3) Expressions as datatypes.  At first blush, Don doesn't
> > think this is
> > illegal (as was discussed in Danvers).  Consequently, I've
> > posted an issue
> > for the RTF on the OMG issues list about this topic so that
> it can be
> > considered more fully.  Don and I will take a more detailed
> > look at it in
> > the near future.
> > ==> Please see the last paragraph on page 56 of MOF 1.3: All
> > (Attributes)
> > have a "type" that is
> > ==> represented using a Data Type ..., and page 57 on
> "type:": This is
> > expressed as the name of
> > ==> a Data Type defined in Section 3.6, or as the name of a
> > CORBA data type
> > ...
> > ==> Therefore, it is illegal to use a CWM-defined Expression
> > (an instance
> > of a MOF Class) as the
> > ==> type in defining a CWM attribute (an instance of a MOF
> attribute).
> > ==> I am copying Sridhar on this note so that he can confirm
> > or rebuke what
> > are stated in MOF 1.3.
> >
> > Doug
> >
> >
> >
> >
> >
> >
> >
>
>
> The information contained in this email and any attached files are
> confidential and intended solely for the addressee(s). The
> e-mail may be
> legally privileged or prohibited from disclosure and unauthorised
use.
> If you are not the named addressee you may not use, copy or
> disclose this
> information to any other person.  If you received this
> message in error
> please notify the sender immediately.
> Any views or opinions presented here may be solely those of
> the originator
> and do not necessarily reflect those of the Company.
>
>
>
>
>


The information contained in this email and any attached files are
confidential and intended solely for the addressee(s). The e-mail may
be
legally privileged or prohibited from disclosure and unauthorised use.
If you are not the named addressee you may not use, copy or disclose
this
information to any other person.  If you received this message in
error
please notify the sender immediately.
Any views or opinions presented here may be solely those of the
originator
and do not necessarily reflect those of the Company.

Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
To: "Pete Rivett" <pete.rivett@adaptive.com>
Cc: cwm-rtf@omg.org, "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>,
   John_Poole@hyperion.com, "'Iyengar, Sridhar'"
   <Sridhar.Iyengar2@unisys.com>
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OF46607F00.0DFF9F87-ON88256A8E.005DBC8D@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Thu, 19 Jul 2001 10:25:55 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
|June 18, 2001) at
 07/19/2001 11:25:56 AM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: CLPe9bl`d9>AEe9hFd!!


Pete,

I hope Sridhar can clarify what MOF 1.3 means.

CWM Expression inherits from CWM Element and is not a CWM Classifier.
Therefore, it cannot be used as the type of a M1 CWM Attribute.

If CWM Expression, being an instance of a MOF Class, can be used as
the
type of a M2 CWM attribute (an instance of a MOF Attribute), then
anything
in CWM (being an instance of a MOF Class) can be used as the type of a
M2
CWM attribute. For example, a CWM Element, a CWM Attribute, a CWM
Generalization, a CWM Association, a CWM AssociationEnd, etc.

Are you sure?

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                           
                    "Pete Rivett"                                                                          
                    <pete.rivett@ada       To:
                    <John_Poole@hyperion.com>                               
                    ptive.com>             cc:     Dan Chang/Santa
                    Teresa/IBM@IBMUS, "'Tolbert, Doug M'"   
                                            <Doug.Tolbert@unisys.com>,
                    <cwm-rtf@omg.org>, "'Iyengar,       
                    07/19/01 08:20          Sridhar'"
                    <Sridhar.Iyengar2@unisys.com>                        
                    AM                     Subject:     RE: "Baisley"
                    questions from Danvers CWM RTF       
                                            meeting                                                        
                                                                                                           
                                                                                                           
                                                                                                           



John,
I hope I made the point in my previous mail that it is NOT illegal to
use
CWM's Expression class as the type of an attribute.

Nonetheless your conditional statement is correct - although for
clarity
I'd
replace "a CWM Expression" (which could be read as "an instance of the
class
Expression") with "CWM's Expression class".

Pete

> -----Original Message-----
> From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com]
> Sent: 19 July 2001 15:48
> To: Pete Rivett
> Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org;
> 'Iyengar, Sridhar'
> Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
>
> Hi all,
>
> It seems to me, by way of analogy, that if it is illegal for
> me to use a
> CWM Expression as an attribute type in another CWM class,
> then it would be
> equally illegal for me to use any other CWM class as the type of an
> object-valued attribute in another CWM class (e.g., ExpressionNode).
>
> Is this a reasonable conclusion, or am I mistaken about this?
>
> - John
>
>
>
>
>
>
> "Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001 10:49:57 PM
>
> To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
>       <Doug.Tolbert@unisys.com>
> cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
> <Sridhar.Iyengar2@unisys.com>
> Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
> Dan,
> Re your comment on point 3):
> The section you quote from MOF 1.3 is in the section about
> the MOF Model
> itself (the M3): it's in effect saying that although the MOF M3 is
an
> instance of itself, it is (voluntarily) restricting itself to
> only defining
> attributes based on Data Types.
> This is placing no such constraint on M2's such as CWM.
> The (M3) class 'Attribute' inherits the reference 'type' from
> TypedElement
> which refers to Classifier. For Attribute the only constraint on
this
> reference is at the TypedElement level where [C-13] states that "An
> Association may not be the type of a TypedElement"
> (implicitly allowing the
> other 2 subtypes of Classifier - Class and DataType).
>
> I'm not sure this addresses the original question since I was
> not involved
> in that discussion at Danvers: but CWM class 'Expression' may
> legally be
> used as the type of a CWM-defined attribute. Whether it's the
> best way of
> modeling the requirement is another question (e.g. it does not allow
a
> (possibly complex or volatile) Expression (e.g. "Sales tax in
> California")
> to be defined only once and referred to from many places,
> minimizing the
> impact of any change).
>
> Pete
>
> Pete Rivett (pete.rivett@adaptive.com)
> Chief Technology Officer, Adaptive Ltd
> Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL, UK
> Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
> http://www.adaptive.com
>
>
>
> > -----Original Message-----
> > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > Sent: 19 July 2001 01:27
> > To: Tolbert, Doug M
> > Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> > Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> >
> > Doug,
> >
> > My comment is embedded below (==>).
> >
> > Regards,  Dan
> >
> > e-business Data Technology and Standard
> > IBM Silicon Valley Laboratory
> > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > Internet:  dtchang@us.ibm.com
> > VM:          IBMUSM50(DTCHANG)
> > Phone:    (408)-463-2319
> >
> --cut--
> >
> >
> > (3) Expressions as datatypes.  At first blush, Don doesn't
> > think this is
> > illegal (as was discussed in Danvers).  Consequently, I've
> > posted an issue
> > for the RTF on the OMG issues list about this topic so that
> it can be
> > considered more fully.  Don and I will take a more detailed
> > look at it in
> > the near future.
> > ==> Please see the last paragraph on page 56 of MOF 1.3: All
> > (Attributes)
> > have a "type" that is
> > ==> represented using a Data Type ..., and page 57 on
> "type:": This is
> > expressed as the name of
> > ==> a Data Type defined in Section 3.6, or as the name of a
> > CORBA data type
> > ...
> > ==> Therefore, it is illegal to use a CWM-defined Expression
> > (an instance
> > of a MOF Class) as the
> > ==> type in defining a CWM attribute (an instance of a MOF
> attribute).
> > ==> I am copying Sridhar on this note so that he can confirm
> > or rebuke what
> > are stated in MOF 1.3.
> >
> > Doug
> >
> >
> >
> >
> >
> >
> >
>
>
> The information contained in this email and any attached files are
> confidential and intended solely for the addressee(s). The
> e-mail may be
> legally privileged or prohibited from disclosure and unauthorised
use.
> If you are not the named addressee you may not use, copy or
> disclose this
> information to any other person.  If you received this
> message in error
> please notify the sender immediately.
> Any views or opinions presented here may be solely those of
> the originator
> and do not necessarily reflect those of the Company.
>
>
>
>
>


The information contained in this email and any attached files are
confidential and intended solely for the addressee(s). The e-mail may
be
legally privileged or prohibited from disclosure and unauthorised use.
If you are not the named addressee you may not use, copy or disclose
this
information to any other person.  If you received this message in
error
please notify the sender immediately.
Any views or opinions presented here may be solely those of the
originator
and do not necessarily reflect those of the Company.


From: "Martin Matula" <martin.matula@czech.sun.com>
To: "Pete Rivett" <pete.rivett@adaptive.com>, "Dan Chang"
<dtchang@us.ibm.com>
Cc: <cwm-rtf@omg.org>, "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>,
   <John_Poole@hyperion.com>,
   "'Iyengar, Sridhar'" <Sridhar.Iyengar2@unisys.com>
References: <OF46607F00.0DFF9F87-ON88256A8E.005DBC8D@boulder.ibm.com>
Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
Date: Thu, 19 Jul 2001 19:58:37 +0200
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.2919.6700
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: ]R9e9Old!!f71!!~5@e9

Hi Dan,
if CWM attribute means "M2 instance of MOF Attribute in your CWM
metamodel"
then Pete is correct. Any instance of MOF Class or MOF DataType can be
referenced as type from any instance of MOF Attribute in your model.
This means, at M2 level you can model an attribute in CWM metamodel to
be of
type CWM Expression. It will allow you to store instances of CWM
expressions
in this attribute at M1 level.
Martin

----- Original Message -----
From: "Dan Chang" <dtchang@us.ibm.com>
To: "Pete Rivett" <pete.rivett@adaptive.com>
Cc: <cwm-rtf@omg.org>; "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>;
<John_Poole@hyperion.com>; "'Iyengar, Sridhar'"
<Sridhar.Iyengar2@unisys.com>
Sent: Thursday, July 19, 2001 7:25 PM
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting


>
> Pete,
>
> I hope Sridhar can clarify what MOF 1.3 means.
>
> CWM Expression inherits from CWM Element and is not a CWM
>Classifier.
> Therefore, it cannot be used as the type of a M1 CWM Attribute.
>
> If CWM Expression, being an instance of a MOF Class, can be used as
>the
> type of a M2 CWM attribute (an instance of a MOF Attribute), then
>anything
> in CWM (being an instance of a MOF Class) can be used as the type of
>a M2
> CWM attribute. For example, a CWM Element, a CWM Attribute, a CWM
> Generalization, a CWM Association, a CWM AssociationEnd, etc.
>
> Are you sure?
>
> Regards,  Dan
>
> e-business Data Technology and Standard
> IBM Silicon Valley Laboratory
> Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> Internet:  dtchang@us.ibm.com
> VM:          IBMUSM50(DTCHANG)
> Phone:    (408)-463-2319
>
>
>
>                     "Pete Rivett"
>                     <pete.rivett@ada       To:
<John_Poole@hyperion.com>
>                     ptive.com>             cc:     Dan Chang/Santa
Teresa/IBM@IBMUS, "'Tolbert, Doug M'"
>                                             <Doug.Tolbert@unisys.com>,
<cwm-rtf@omg.org>, "'Iyengar,
>                     07/19/01 08:20          Sridhar'"
<Sridhar.Iyengar2@unisys.com>
>                     AM                     Subject:     RE:
>"Baisley"
questions from Danvers CWM RTF
>                                             meeting
>
>
>
>
>
>
> John,
> I hope I made the point in my previous mail that it is NOT illegal
>to use
> CWM's Expression class as the type of an attribute.
>
> Nonetheless your conditional statement is correct - although for
>clarity
> I'd
> replace "a CWM Expression" (which could be read as "an instance of
>the
> class
> Expression") with "CWM's Expression class".
>
> Pete
>
> > -----Original Message-----
> > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com]
> > Sent: 19 July 2001 15:48
> > To: Pete Rivett
> > Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org;
> > 'Iyengar, Sridhar'
> > Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> >
> >
> > Hi all,
> >
> > It seems to me, by way of analogy, that if it is illegal for
> > me to use a
> > CWM Expression as an attribute type in another CWM class,
> > then it would be
> > equally illegal for me to use any other CWM class as the type of
>an
> > object-valued attribute in another CWM class (e.g.,
>ExpressionNode).
> >
> > Is this a reasonable conclusion, or am I mistaken about this?
> >
> > - John
> >
> >
> >
> >
> >
> >
> > "Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001 10:49:57 PM
> >
> > To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
> >       <Doug.Tolbert@unisys.com>
> > cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
> > <Sridhar.Iyengar2@unisys.com>
> > Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> > Dan,
> > Re your comment on point 3):
> > The section you quote from MOF 1.3 is in the section about
> > the MOF Model
> > itself (the M3): it's in effect saying that although the MOF M3 is
>an
> > instance of itself, it is (voluntarily) restricting itself to
> > only defining
> > attributes based on Data Types.
> > This is placing no such constraint on M2's such as CWM.
> > The (M3) class 'Attribute' inherits the reference 'type' from
> > TypedElement
> > which refers to Classifier. For Attribute the only constraint on
>this
> > reference is at the TypedElement level where [C-13] states that
>"An
> > Association may not be the type of a TypedElement"
> > (implicitly allowing the
> > other 2 subtypes of Classifier - Class and DataType).
> >
> > I'm not sure this addresses the original question since I was
> > not involved
> > in that discussion at Danvers: but CWM class 'Expression' may
> > legally be
> > used as the type of a CWM-defined attribute. Whether it's the
> > best way of
> > modeling the requirement is another question (e.g. it does not
>allow a
> > (possibly complex or volatile) Expression (e.g. "Sales tax in
> > California")
> > to be defined only once and referred to from many places,
> > minimizing the
> > impact of any change).
> >
> > Pete
> >
> > Pete Rivett (pete.rivett@adaptive.com)
> > Chief Technology Officer, Adaptive Ltd
> > Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL, UK
> > Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
> > http://www.adaptive.com
> >
> >
> >
> > > -----Original Message-----
> > > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > > Sent: 19 July 2001 01:27
> > > To: Tolbert, Doug M
> > > Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> > > Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> > >
> > >
> > >
> > > Doug,
> > >
> > > My comment is embedded below (==>).
> > >
> > > Regards,  Dan
> > >
> > > e-business Data Technology and Standard
> > > IBM Silicon Valley Laboratory
> > > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > > Internet:  dtchang@us.ibm.com
> > > VM:          IBMUSM50(DTCHANG)
> > > Phone:    (408)-463-2319
> > >
> > --cut--
> > >
> > >
> > > (3) Expressions as datatypes.  At first blush, Don doesn't
> > > think this is
> > > illegal (as was discussed in Danvers).  Consequently, I've
> > > posted an issue
> > > for the RTF on the OMG issues list about this topic so that
> > it can be
> > > considered more fully.  Don and I will take a more detailed
> > > look at it in
> > > the near future.
> > > ==> Please see the last paragraph on page 56 of MOF 1.3: All
> > > (Attributes)
> > > have a "type" that is
> > > ==> represented using a Data Type ..., and page 57 on
> > "type:": This is
> > > expressed as the name of
> > > ==> a Data Type defined in Section 3.6, or as the name of a
> > > CORBA data type
> > > ...
> > > ==> Therefore, it is illegal to use a CWM-defined Expression
> > > (an instance
> > > of a MOF Class) as the
> > > ==> type in defining a CWM attribute (an instance of a MOF
> > attribute).
> > > ==> I am copying Sridhar on this note so that he can confirm
> > > or rebuke what
> > > are stated in MOF 1.3.
> > >
> > > Doug
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
> >
> > The information contained in this email and any attached files are
> > confidential and intended solely for the addressee(s). The
> > e-mail may be
> > legally privileged or prohibited from disclosure and unauthorised
>use.
> > If you are not the named addressee you may not use, copy or
> > disclose this
> > information to any other person.  If you received this
> > message in error
> > please notify the sender immediately.
> > Any views or opinions presented here may be solely those of
> > the originator
> > and do not necessarily reflect those of the Company.
> >
> >
> >
> >
> >
>
>
> The information contained in this email and any attached files are
> confidential and intended solely for the addressee(s). The e-mail
>may be
> legally privileged or prohibited from disclosure and unauthorised
>use.
> If you are not the named addressee you may not use, copy or disclose
>this
> information to any other person.  If you received this message in
>error
> please notify the sender immediately.
> Any views or opinions presented here may be solely those of the
>originator
> and do not necessarily reflect those of the Company.
>
>
>
>
Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
To: "Martin Matula" <martin.matula@czech.sun.com>
Cc: cwm-rtf@omg.org, "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>,
   John_Poole@hyperion.com, "Pete Rivett" <pete.rivett@adaptive.com>,
   "'Iyengar, Sridhar'" <Sridhar.Iyengar2@unisys.com>
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OFA49F86C5.4AB46859-ON88256A8E.0063E234@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Thu, 19 Jul 2001 11:58:21 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
>|June 18, 2001) at
 07/19/2001 12:58:30 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: Z`R!!$Mnd9X*2e90>(!!


Martin,

If so, then we need to re-examine CWM.

For example, the CWM Object Model is meant to be a subset of
UML. However,
in UML, Expression is a Data Type and, in the IDL binding, it maps to
a
struct. In CWM, Expression is not a Data Type and, in the IDL binding,
it
maps to an interface. Also, in UML, all Expresions (Expression,
BooleanExpression, ProcedureExpression, ...) have the same
"attributes"
(language and body), map to the identical struct, and cannot be
specialized
as normal classes. In CWM, Expressions (Expression, BooleanExpression,
ProcudureExpression) can have different attributes, map to different
interfaces, and can be specialized as normal classes.

We definitely did not intend to make such modeling changes (on
Expressions,
Multiplicity, MultiplicityRange, etc.) when we derived the CWM Object
Model.

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                              
                    "Martin Matula"                                                                           
                    <martin.matula@czec       To:     "Pete Rivett"
                    <pete.rivett@adaptive.com>, Dan           
                    h.sun.com>                 Chang/Santa
                    Teresa/IBM@IBMUS                                   
                                              cc:
                    <cwm-rtf@omg.org>, "'Tolbert, Doug M'"                  
                    07/19/01 10:58 AM
                    <Doug.Tolbert@unisys.com>,
                    <John_Poole@hyperion.com>,          
                                               "'Iyengar, Sridhar'"
                    <Sridhar.Iyengar2@unisys.com>             
                                              Subject:     Re:
                    "Baisley" questions from Danvers CWM RTF       
                                               meeting                                                        
                                                                                                              
                                                                                                              
                                                                                                              



Hi Dan,
if CWM attribute means "M2 instance of MOF Attribute in your CWM
metamodel"
then Pete is correct. Any instance of MOF Class or MOF DataType can be
referenced as type from any instance of MOF Attribute in your model.
This means, at M2 level you can model an attribute in CWM metamodel to
be
of
type CWM Expression. It will allow you to store instances of CWM
expressions
in this attribute at M1 level.
Martin

----- Original Message -----
From: "Dan Chang" <dtchang@us.ibm.com>
To: "Pete Rivett" <pete.rivett@adaptive.com>
Cc: <cwm-rtf@omg.org>; "'Tolbert, Doug M'" <Doug.Tolbert@unisys.com>;
<John_Poole@hyperion.com>; "'Iyengar, Sridhar'"
<Sridhar.Iyengar2@unisys.com>
Sent: Thursday, July 19, 2001 7:25 PM
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting


>
> Pete,
>
> I hope Sridhar can clarify what MOF 1.3 means.
>
> CWM Expression inherits from CWM Element and is not a CWM
>Classifier.
> Therefore, it cannot be used as the type of a M1 CWM Attribute.
>
> If CWM Expression, being an instance of a MOF Class, can be used as
>the

From: "Baisley, Donald E" <Donald.Baisley@unisys.com>
To: cwm-rtf@omg.org
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Date: Thu, 19 Jul 2001 19:59:55 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: A3B!!/b7e9Yp4e9=@Xd9
Status: RO

Hi,

Since my name is in the subject of this email thread I thought I
should
interject something.  As Pete has pointed out, section 3.2 of the MOF
Specification does not prevent a class being the type of an attribute.
Section 3.2 merely describes how the MOF Model is described in the
specification.  It just happens that MOF Model does not ever use a
class as
the type of an attribute, so the wording refers to data types.

MOF explicitly allows an attribute's type to be a class.  The UML 1.4
metamodel uses this capability with Expression and Multiplicity, both
of
which are classes (not data types) in UML 1.4.  According to MOF
rules,
there are composition semantics when an attribute's type is a class.
For
example, if an expression is assigned as the default value of a
parameter,
then that expression belongs to the parameter and cannot be owned by
any
other object (just as if a composite association was used between
parameter
and expression).

I hope this clears things up.

Regards,
Don

-----Original Message-----
From: Iyengar, Sridhar 
Sent: Thursday, July 19, 2001 12:03 PM
To: Don Baisley; Doug Tolbert
Subject: FW: "Baisley" questions from Danvers CWM RTF meeting
Importance: High


Can we calm down Dan - what is his problem?

------------------------------------------------------------------------
----
Sridhar Iyengar
Unisys Fellow, Director of Advanced Technology
25725, Jeronimo Road
Mission Viejo, CA 92691

E-mail : Sridhar.iyengar2@unisys.com
Phone : 949-380-5692
Fax     : 949-380-6632
------------------------------------------------------------------------
--------------




> -----Original Message-----
> From: Dan Chang [mailto:dtchang@us.ibm.com] 
> Sent: Thursday, July 19, 2001 11:58 AM
> To: Martin Matula
> Cc: cwm-rtf@omg.org; 'Tolbert, Doug M'; John_Poole@hyperion.com;
Pete
> Rivett; 'Iyengar, Sridhar'
> Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> 
> 
> 
> Martin,
> 
> If so, then we need to re-examine CWM.
> 
> For example, the CWM Object Model is meant to be a subset of 
> UML. However,
> in UML, Expression is a Data Type and, in the IDL binding, it 
> maps to a
> struct. In CWM, Expression is not a Data Type and, in the IDL 
> binding, it
> maps to an interface. Also, in UML, all Expresions (Expression,
> BooleanExpression, ProcedureExpression, ...) have the same 
> "attributes"
> (language and body), map to the identical struct, and cannot 
> be specialized
> as normal classes. In CWM, Expressions (Expression,
BooleanExpression,
> ProcudureExpression) can have different attributes, map to different
> interfaces, and can be specialized as normal classes.
> 
> We definitely did not intend to make such modeling changes 
> (on Expressions,
> Multiplicity, MultiplicityRange, etc.) when we derived the CWM
Object
> Model.
> 
> Regards,  Dan
> 
> e-business Data Technology and Standard
> IBM Silicon Valley Laboratory
> Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> Internet:  dtchang@us.ibm.com
> VM:          IBMUSM50(DTCHANG)
> Phone:    (408)-463-2319
> 
> 
>                                                               
>                                                 
>                     "Martin Matula"                           
>                                                 
>                     <martin.matula@czec       To:     "Pete 
> Rivett" <pete.rivett@adaptive.com>, Dan           
>                     h.sun.com>                 Chang/Santa 
> Teresa/IBM@IBMUS                                   
>                                               cc:     
> <cwm-rtf@omg.org>, "'Tolbert, Doug M'"                  
>                     07/19/01 10:58 AM          
> <Doug.Tolbert@unisys.com>, <John_Poole@hyperion.com>,          
>                                                "'Iyengar, 
> Sridhar'" <Sridhar.Iyengar2@unisys.com>             
>                                               Subject:     
> Re: "Baisley" questions from Danvers CWM RTF       
>                                                meeting        
>                                                 
>                                                               
>                                                 
>                                                               
>                                                 
>                                                               
>                                                 
> 
> 
> 
> Hi Dan,
> if CWM attribute means "M2 instance of MOF Attribute in your 
> CWM metamodel"
> then Pete is correct. Any instance of MOF Class or MOF DataType can
be
> referenced as type from any instance of MOF Attribute in your model.
> This means, at M2 level you can model an attribute in CWM 
> metamodel to be
> of
> type CWM Expression. It will allow you to store instances of CWM
> expressions
> in this attribute at M1 level.
> Martin
> 
> ----- Original Message -----
> From: "Dan Chang" <dtchang@us.ibm.com>
> To: "Pete Rivett" <pete.rivett@adaptive.com>
> Cc: <cwm-rtf@omg.org>; "'Tolbert, Doug M'"
<Doug.Tolbert@unisys.com>;
> <John_Poole@hyperion.com>; "'Iyengar, Sridhar'"
> <Sridhar.Iyengar2@unisys.com>
> Sent: Thursday, July 19, 2001 7:25 PM
> Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
> 
> 
> >
> > Pete,
> >
> > I hope Sridhar can clarify what MOF 1.3 means.
> >
> > CWM Expression inherits from CWM Element and is not a CWM 
> Classifier.
> > Therefore, it cannot be used as the type of a M1 CWM Attribute.
> >
> > If CWM Expression, being an instance of a MOF Class, can be 
> used as the
> > type of a M2 CWM attribute (an instance of a MOF Attribute), then
> anything
> > in CWM (being an instance of a MOF Class) can be used as 
> the type of a M2
> > CWM attribute. For example, a CWM Element, a CWM Attribute, a CWM
> > Generalization, a CWM Association, a CWM AssociationEnd, etc.
> >
> > Are you sure?
> >
> > Regards,  Dan
> >
> > e-business Data Technology and Standard
> > IBM Silicon Valley Laboratory
> > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > Internet:  dtchang@us.ibm.com
> > VM:          IBMUSM50(DTCHANG)
> > Phone:    (408)-463-2319
> >
> >
> >
> >                     "Pete Rivett"
> >                     <pete.rivett@ada       To:
> <John_Poole@hyperion.com>
> >                     ptive.com>             cc:     Dan Chang/Santa
> Teresa/IBM@IBMUS, "'Tolbert, Doug M'"
> >                                             
> <Doug.Tolbert@unisys.com>,
> <cwm-rtf@omg.org>, "'Iyengar,
> >                     07/19/01 08:20          Sridhar'"
> <Sridhar.Iyengar2@unisys.com>
> >                     AM                     Subject:     RE: 
> "Baisley"
> questions from Danvers CWM RTF
> >                                             meeting
> >
> >
> >
> >
> >
> >
> > John,
> > I hope I made the point in my previous mail that it is NOT 
> illegal to use
> > CWM's Expression class as the type of an attribute.
> >
> > Nonetheless your conditional statement is correct - 
> although for clarity
> > I'd
> > replace "a CWM Expression" (which could be read as "an 
> instance of the
> > class
> > Expression") with "CWM's Expression class".
> >
> > Pete
> >
> > > -----Original Message-----
> > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com]
> > > Sent: 19 July 2001 15:48
> > > To: Pete Rivett
> > > Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org;
> > > 'Iyengar, Sridhar'
> > > Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
> > >
> > >
> > >
> > >
> > > Hi all,
> > >
> > > It seems to me, by way of analogy, that if it is illegal for
> > > me to use a
> > > CWM Expression as an attribute type in another CWM class,
> > > then it would be
> > > equally illegal for me to use any other CWM class as the 
> type of an
> > > object-valued attribute in another CWM class (e.g., 
> ExpressionNode).
> > >
> > > Is this a reasonable conclusion, or am I mistaken about this?
> > >
> > > - John
> > >
> > >
> > >
> > >
> > >
> > >
> > > "Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001 10:49:57
PM
> > >
> > > To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
> > >       <Doug.Tolbert@unisys.com>
> > > cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
> > > <Sridhar.Iyengar2@unisys.com>
> > > Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting
> > >
> > >
> > > Dan,
> > > Re your comment on point 3):
> > > The section you quote from MOF 1.3 is in the section about
> > > the MOF Model
> > > itself (the M3): it's in effect saying that although the 
> MOF M3 is an
> > > instance of itself, it is (voluntarily) restricting itself to
> > > only defining
> > > attributes based on Data Types.
> > > This is placing no such constraint on M2's such as CWM.
> > > The (M3) class 'Attribute' inherits the reference 'type' from
> > > TypedElement
> > > which refers to Classifier. For Attribute the only 
> constraint on this
> > > reference is at the TypedElement level where [C-13] 
> states that "An
> > > Association may not be the type of a TypedElement"
> > > (implicitly allowing the
> > > other 2 subtypes of Classifier - Class and DataType).
> > >
> > > I'm not sure this addresses the original question since I was
> > > not involved
> > > in that discussion at Danvers: but CWM class 'Expression' may
> > > legally be
> > > used as the type of a CWM-defined attribute. Whether it's the
> > > best way of
> > > modeling the requirement is another question (e.g. it 
> does not allow a
> > > (possibly complex or volatile) Expression (e.g. "Sales tax in
> > > California")
> > > to be defined only once and referred to from many places,
> > > minimizing the
> > > impact of any change).
> > >
> > > Pete
> > >
> > > Pete Rivett (pete.rivett@adaptive.com)
> > > Chief Technology Officer, Adaptive Ltd
> > > Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL,
UK
> > > Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
> > > http://www.adaptive.com
> > >
> > >
> > >
> > > > -----Original Message-----
> > > > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > > > Sent: 19 July 2001 01:27
> > > > To: Tolbert, Doug M
> > > > Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> > > > Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> > > >
> > > >
> > > >
> > > > Doug,
> > > >
> > > > My comment is embedded below (==>).
> > > >
> > > > Regards,  Dan
> > > >
> > > > e-business Data Technology and Standard
> > > > IBM Silicon Valley Laboratory
> > > > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > > > Internet:  dtchang@us.ibm.com
> > > > VM:          IBMUSM50(DTCHANG)
> > > > Phone:    (408)-463-2319
> > > >
> > > --cut--
> > > >
> > > >
> > > > (3) Expressions as datatypes.  At first blush, Don doesn't
> > > > think this is
> > > > illegal (as was discussed in Danvers).  Consequently, I've
> > > > posted an issue
> > > > for the RTF on the OMG issues list about this topic so that
> > > it can be
> > > > considered more fully.  Don and I will take a more detailed
> > > > look at it in
> > > > the near future.
> > > > ==> Please see the last paragraph on page 56 of MOF 1.3: All
> > > > (Attributes)
> > > > have a "type" that is
> > > > ==> represented using a Data Type ..., and page 57 on
> > > "type:": This is
> > > > expressed as the name of
> > > > ==> a Data Type defined in Section 3.6, or as the name of a
> > > > CORBA data type
> > > > ...
> > > > ==> Therefore, it is illegal to use a CWM-defined Expression
> > > > (an instance
> > > > of a MOF Class) as the
> > > > ==> type in defining a CWM attribute (an instance of a MOF
> > > attribute).
> > > > ==> I am copying Sridhar on this note so that he can confirm
> > > > or rebuke what
> > > > are stated in MOF 1.3.
> > > >
> > > > Doug
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > The information contained in this email and any attached files
are
> > > confidential and intended solely for the addressee(s). The
> > > e-mail may be
> > > legally privileged or prohibited from disclosure and 
> unauthorised use.
> > > If you are not the named addressee you may not use, copy or
> > > disclose this
> > > information to any other person.  If you received this
> > > message in error
> > > please notify the sender immediately.
> > > Any views or opinions presented here may be solely those of
> > > the originator
> > > and do not necessarily reflect those of the Company.
> > >
> > >
> > >
> > >
> > >
> >
> >
> > The information contained in this email and any attached files are
> > confidential and intended solely for the addressee(s). The 
> e-mail may be
> > legally privileged or prohibited from disclosure and 
> unauthorised use.
> > If you are not the named addressee you may not use, copy or 
> disclose this
> > information to any other person.  If you received this 
> message in error
> > please notify the sender immediately.
> > Any views or opinions presented here may be solely those of the
> originator
> > and do not necessarily reflect those of the Company.
> >
> >
> >
> >
> 
> 
> 
> 
> 
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
To: "Baisley, Donald E" <Donald.Baisley@unisys.com>
Cc: cwm-rtf@omg.org
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OFD22B1D84.97814394-ON88256A8F.0058B3D5@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Fri, 20 Jul 2001 09:45:31 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
|June 18, 2001) at
 07/20/2001 10:45:32 AM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: 4GL!!5d>!!3;O!!Qm/!!


Don,

Unfortunately, it does not.

First, in the draft UML 1.4 spec dated 1/01 that I have, Expression,
etc.
are still in the Data Type package. Nothing says that they are
different
from UML 1.3. If you have a newer version that is different, please
send it
to me or the CWM RTF (I could not download the 2/01 version from the
OMG
web site)..

Second, whether Expression ,etc. are data types (mapped to struct or
equivalent) or classes (mapped to interface/class) has implication on
modeling and usage/performance. For example, the CWM ExpressionNode
inherits from CWM Element and not CWM Expression. The assumption was
that
CWM Expression, as Expression in UML 1.3, cannot be specialized like a
normal class.

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                            
                    "Baisley, Donald                                                                        
                    E"                      To:     cwm-rtf@omg.org                                         
                    <Donald.Baisley@u       cc:                                                             
                    nisys.com>              Subject:     RE: "Baisley"
                    questions from Danvers CWM RTF       
                                             meeting                                                        
                    07/19/01 05:59 PM                                                                       
                                                                                                            
                                                                                                            



Hi,

Since my name is in the subject of this email thread I thought I
should
interject something.  As Pete has pointed out, section 3.2 of the MOF
Specification does not prevent a class being the type of an attribute.
Section 3.2 merely describes how the MOF Model is described in the
specification.  It just happens that MOF Model does not ever use a
class as
the type of an attribute, so the wording refers to data types.

MOF explicitly allows an attribute's type to be a class.  The UML 1.4
metamodel uses this capability with Expression and Multiplicity, both
of
which are classes (not data types) in UML 1.4.  According to MOF
rules,
there are composition semantics when an attribute's type is a class.
For
example, if an expression is assigned as the default value of a
parameter,
then that expression belongs to the parameter and cannot be owned by
any
other object (just as if a composite association was used between
parameter
and expression).

I hope this clears things up.

Regards,
Don

-----Original Message-----
From: Iyengar, Sridhar
Sent: Thursday, July 19, 2001 12:03 PM
To: Don Baisley; Doug Tolbert
Subject: FW: "Baisley" questions from Danvers CWM RTF meeting
Importance: High


Can we calm down Dan - what is his problem?

------------------------------------------------------------------------
----
Sridhar Iyengar
Unisys Fellow, Director of Advanced Technology
25725, Jeronimo Road
Mission Viejo, CA 92691

E-mail : Sridhar.iyengar2@unisys.com
Phone : 949-380-5692
Fax     : 949-380-6632
------------------------------------------------------------------------
--------------




> -----Original Message-----
> From: Dan Chang [mailto:dtchang@us.ibm.com]
> Sent: Thursday, July 19, 2001 11:58 AM
> To: Martin Matula
> Cc: cwm-rtf@omg.org; 'Tolbert, Doug M'; John_Poole@hyperion.com;
Pete
> Rivett; 'Iyengar, Sridhar'
> Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
> Martin,
>
> If so, then we need to re-examine CWM.
>
> For example, the CWM Object Model is meant to be a subset of
> UML. However,
> in UML, Expression is a Data Type and, in the IDL binding, it
> maps to a
> struct. In CWM, Expression is not a Data Type and, in the IDL
> binding, it
> maps to an interface. Also, in UML, all Expresions (Expression,
> BooleanExpression, ProcedureExpression, ...) have the same
> "attributes"
> (language and body), map to the identical struct, and cannot
> be specialized
> as normal classes. In CWM, Expressions (Expression,
BooleanExpression,
> ProcudureExpression) can have different attributes, map to different
> interfaces, and can be specialized as normal classes.
>
> We definitely did not intend to make such modeling changes
> (on Expressions,
> Multiplicity, MultiplicityRange, etc.) when we derived the CWM
Object

From: "Martin Matula" <martin.matula@czech.sun.com>
To: "Baisley, Donald E" <Donald.Baisley@unisys.com>,
   "Dan Chang" <dtchang@us.ibm.com>
Cc: <cwm-rtf@omg.org>
References: <OFD22B1D84.97814394-ON88256A8F.0058B3D5@boulder.ibm.com>
Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
Date: Fri, 20 Jul 2001 20:32:16 +0200
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.2919.6700
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: (lNe9BpHe92MHe97(%!!

Hi Dan,
the fact that something is defined in UML Data Type package does not
automatically mean that it is an instance of MOF DataType. The truth
is,
that Expression is really instance of MOF Class rather than instance
of MOF
DataType. If it wasn't, TimeExpression and other classes would not be
able
to inherit from it (MOF does not allow inheritance of DataTypes).
Martin

----- Original Message -----
From: "Dan Chang" <dtchang@us.ibm.com>
To: "Baisley, Donald E" <Donald.Baisley@unisys.com>
Cc: <cwm-rtf@omg.org>
Sent: Friday, July 20, 2001 6:45 PM
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting


>
> Don,
>
> Unfortunately, it does not.
>
> First, in the draft UML 1.4 spec dated 1/01 that I have, Expression,
>etc.
> are still in the Data Type package. Nothing says that they are
>different
> from UML 1.3. If you have a newer version that is different, please
>send
it
> to me or the CWM RTF (I could not download the 2/01 version from the
>OMG
> web site)..
>
> Second, whether Expression ,etc. are data types (mapped to struct or
> equivalent) or classes (mapped to interface/class) has implication
>on
> modeling and usage/performance. For example, the CWM ExpressionNode
> inherits from CWM Element and not CWM Expression. The assumption was
>that
> CWM Expression, as Expression in UML 1.3, cannot be specialized like
>a
> normal class.
>
> Regards,  Dan
>
> e-business Data Technology and Standard
> IBM Silicon Valley Laboratory
> Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> Internet:  dtchang@us.ibm.com
> VM:          IBMUSM50(DTCHANG)
> Phone:    (408)-463-2319

From: "Pete Rivett" <pete.rivett@adaptive.com>
To: "'Dan Chang'" <dtchang@us.ibm.com>,
   "'Baisley, Donald E'" <Donald.Baisley@unisys.com>
Cc: <cwm-rtf@omg.org>
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Date: Fri, 20 Jul 2001 19:38:22 +0100
Message-ID: <003501c1114b$288811b0$114c04c8@CHIMAY>
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2911.0)
Importance: Normal
In-Reply-To: <OFD22B1D84.97814394-ON88256A8F.0058B3D5@boulder.ibm.com>
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.3018.1300
Content-Type: text/plain;
	      charset="us-ascii"
X-UIDL: 2*K!!:Nf!!T-@!!!YRd9

Dan,
Expression is in the UML "Datatype" package but "datatype" is here
just a
package name: it's not necessarily the case that the contents of this
package must all instances of MOF Datatype.

Expression is indeed an instance of MOF Class not MOF DataType: the
way you
can tell is that the 'real' MOF datatypes have the <<datatype>>
stereotype:
Multiplicity and Expression do not. Also because Expression has lots
of
subtypes - which are not allowed for MOF Datatypes (MOF constraint
[C-19]).
If you look at the UML 1.4 XMI file, Multiplicity and Expression are
present
as instances of Model.Class, and in the IDL as Interfaces. I'll email
you
the files privately since sending files to OMG mailing lists is not
permitted.
You're right that this seems an unheralded change from UML 1.3.

In another email you pointed out that:
"... in UML, which is M2, only data types defined in the Data Type
package
can be used to declare
the types of class attributes." I quickly looked in the UML 1.4 spec
but
could not find this restriction: I'd be interested to discover whether
this
is a Constraint in UML or a convention/policy that has been adopted?

Pete

> -----Original Message-----
> From: Dan Chang [mailto:dtchang@us.ibm.com]
> Sent: 20 July 2001 17:46
> To: Baisley, Donald E
> Cc: cwm-rtf@omg.org
> Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
> Don,
>
> Unfortunately, it does not.
>
> First, in the draft UML 1.4 spec dated 1/01 that I have,
> Expression, etc.
> are still in the Data Type package. Nothing says that they
> are different
> from UML 1.3. If you have a newer version that is different,
> please send it
> to me or the CWM RTF (I could not download the 2/01 version
> from the OMG
> web site)..
>
> Second, whether Expression ,etc. are data types (mapped to struct or
> equivalent) or classes (mapped to interface/class) has implication
on
> modeling and usage/performance. For example, the CWM ExpressionNode
> inherits from CWM Element and not CWM Expression. The
> assumption was that
> CWM Expression, as Expression in UML 1.3, cannot be specialized like
a
> normal class.
>
> Regards,  Dan
>
> e-business Data Technology and Standard
> IBM Silicon Valley Laboratory
> Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> Internet:  dtchang@us.ibm.com
> VM:          IBMUSM50(DTCHANG)
> Phone:    (408)-463-2319
>
>
>
>
>                     "Baisley, Donald
>
>                     E"                      To:
> cwm-rtf@omg.org
>                     <Donald.Baisley@u       cc:
>
>                     nisys.com>              Subject:     RE:
> "Baisley" questions from Danvers CWM RTF
>                                              meeting
>
>                     07/19/01 05:59 PM
>
>
>
>
>
>
>
>
> Hi,
>
> Since my name is in the subject of this email thread I
> thought I should
> interject something.  As Pete has pointed out, section 3.2 of the
MOF
> Specification does not prevent a class being the type of an
attribute.
> Section 3.2 merely describes how the MOF Model is described in the
> specification.  It just happens that MOF Model does not ever
> use a class as
> the type of an attribute, so the wording refers to data types.
>
> MOF explicitly allows an attribute's type to be a class.  The UML
1.4
> metamodel uses this capability with Expression and
> Multiplicity, both of
> which are classes (not data types) in UML 1.4.  According to
> MOF rules,
> there are composition semantics when an attribute's type is a
> class.  For
> example, if an expression is assigned as the default value of
> a parameter,
> then that expression belongs to the parameter and cannot be
> owned by any
> other object (just as if a composite association was used
> between parameter
> and expression).
>
> I hope this clears things up.
>
> Regards,
> Don
>
> -----Original Message-----
> From: Iyengar, Sridhar
> Sent: Thursday, July 19, 2001 12:03 PM
> To: Don Baisley; Doug Tolbert
> Subject: FW: "Baisley" questions from Danvers CWM RTF meeting
> Importance: High
>
>
> Can we calm down Dan - what is his problem?
>
> --------------------------------------------------------------
> ----------
> ----
> Sridhar Iyengar
> Unisys Fellow, Director of Advanced Technology
> 25725, Jeronimo Road
> Mission Viejo, CA 92691
>
> E-mail : Sridhar.iyengar2@unisys.com
> Phone : 949-380-5692
> Fax     : 949-380-6632
> --------------------------------------------------------------
> ----------
> --------------
>
>
>
>
> > -----Original Message-----
> > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > Sent: Thursday, July 19, 2001 11:58 AM
> > To: Martin Matula
> > Cc: cwm-rtf@omg.org; 'Tolbert, Doug M';
> John_Poole@hyperion.com; Pete
> > Rivett; 'Iyengar, Sridhar'
> > Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> >
> > Martin,
> >
> > If so, then we need to re-examine CWM.
> >
> > For example, the CWM Object Model is meant to be a subset of
> > UML. However,
> > in UML, Expression is a Data Type and, in the IDL binding, it
> > maps to a
> > struct. In CWM, Expression is not a Data Type and, in the IDL
> > binding, it
> > maps to an interface. Also, in UML, all Expresions (Expression,
> > BooleanExpression, ProcedureExpression, ...) have the same
> > "attributes"
> > (language and body), map to the identical struct, and cannot
> > be specialized
> > as normal classes. In CWM, Expressions (Expression,
> BooleanExpression,
> > ProcudureExpression) can have different attributes, map to
different
> > interfaces, and can be specialized as normal classes.
> >
> > We definitely did not intend to make such modeling changes
> > (on Expressions,
> > Multiplicity, MultiplicityRange, etc.) when we derived the
> CWM Object
> > Model.
> >
> > Regards,  Dan
> >
> > e-business Data Technology and Standard
> > IBM Silicon Valley Laboratory
> > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > Internet:  dtchang@us.ibm.com
> > VM:          IBMUSM50(DTCHANG)
> > Phone:    (408)-463-2319
> >
> >
> >
> >
> >                     "Martin Matula"
> >
> >                     <martin.matula@czec       To:     "Pete
> > Rivett" <pete.rivett@adaptive.com>, Dan
> >                     h.sun.com>                 Chang/Santa
> > Teresa/IBM@IBMUS
> >                                               cc:
> > <cwm-rtf@omg.org>, "'Tolbert, Doug M'"
> >                     07/19/01 10:58 AM
> > <Doug.Tolbert@unisys.com>, <John_Poole@hyperion.com>,
> >                                                "'Iyengar,
> > Sridhar'" <Sridhar.Iyengar2@unisys.com>
> >                                               Subject:
> > Re: "Baisley" questions from Danvers CWM RTF
> >                                                meeting
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > Hi Dan,
> > if CWM attribute means "M2 instance of MOF Attribute in your
> > CWM metamodel"
> > then Pete is correct. Any instance of MOF Class or MOF
> DataType can be
> > referenced as type from any instance of MOF Attribute in your
model.
> > This means, at M2 level you can model an attribute in CWM
> > metamodel to be
> > of
> > type CWM Expression. It will allow you to store instances of CWM
> > expressions
> > in this attribute at M1 level.
> > Martin
> >
> > ----- Original Message -----
> > From: "Dan Chang" <dtchang@us.ibm.com>
> > To: "Pete Rivett" <pete.rivett@adaptive.com>
> > Cc: <cwm-rtf@omg.org>; "'Tolbert, Doug M'"
> <Doug.Tolbert@unisys.com>;
> > <John_Poole@hyperion.com>; "'Iyengar, Sridhar'"
> > <Sridhar.Iyengar2@unisys.com>
> > Sent: Thursday, July 19, 2001 7:25 PM
> > Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> > >
> > > Pete,
> > >
> > > I hope Sridhar can clarify what MOF 1.3 means.
> > >
> > > CWM Expression inherits from CWM Element and is not a CWM
> > Classifier.
> > > Therefore, it cannot be used as the type of a M1 CWM Attribute.
> > >
> > > If CWM Expression, being an instance of a MOF Class, can be
> > used as the
> > > type of a M2 CWM attribute (an instance of a MOF Attribute),
then
> > anything
> > > in CWM (being an instance of a MOF Class) can be used as
> > the type of a M2
> > > CWM attribute. For example, a CWM Element, a CWM Attribute, a
CWM
> > > Generalization, a CWM Association, a CWM AssociationEnd, etc.
> > >
> > > Are you sure?
> > >
> > > Regards,  Dan
> > >
> > > e-business Data Technology and Standard
> > > IBM Silicon Valley Laboratory
> > > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > > Internet:  dtchang@us.ibm.com
> > > VM:          IBMUSM50(DTCHANG)
> > > Phone:    (408)-463-2319
> > >
> > >
> > >
> > >                     "Pete Rivett"
> > >                     <pete.rivett@ada       To:
> > <John_Poole@hyperion.com>
> > >                     ptive.com>             cc:     Dan
Chang/Santa
> > Teresa/IBM@IBMUS, "'Tolbert, Doug M'"
> > >
> > <Doug.Tolbert@unisys.com>,
> > <cwm-rtf@omg.org>, "'Iyengar,
> > >                     07/19/01 08:20          Sridhar'"
> > <Sridhar.Iyengar2@unisys.com>
> > >                     AM                     Subject:     RE:
> > "Baisley"
> > questions from Danvers CWM RTF
> > >                                             meeting
> > >
> > >
> > >
> > >
> > >
> > >
> > > John,
> > > I hope I made the point in my previous mail that it is NOT
> > illegal to use
> > > CWM's Expression class as the type of an attribute.
> > >
> > > Nonetheless your conditional statement is correct -
> > although for clarity
> > > I'd
> > > replace "a CWM Expression" (which could be read as "an
> > instance of the
> > > class
> > > Expression") with "CWM's Expression class".
> > >
> > > Pete
> > >
> > > > -----Original Message-----
> > > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com]
> > > > Sent: 19 July 2001 15:48
> > > > To: Pete Rivett
> > > > Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org;
> > > > 'Iyengar, Sridhar'
> > > > Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
> > > >
> > > >
> > > >
> > > >
> > > > Hi all,
> > > >
> > > > It seems to me, by way of analogy, that if it is illegal for
> > > > me to use a
> > > > CWM Expression as an attribute type in another CWM class,
> > > > then it would be
> > > > equally illegal for me to use any other CWM class as the
> > type of an
> > > > object-valued attribute in another CWM class (e.g.,
> > ExpressionNode).
> > > >
> > > > Is this a reasonable conclusion, or am I mistaken about this?
> > > >
> > > > - John
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > "Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001
> 10:49:57 PM
> > > >
> > > > To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
> > > >       <Doug.Tolbert@unisys.com>
> > > > cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
> > > > <Sridhar.Iyengar2@unisys.com>
> > > > Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting
> > > >
> > > >
> > > > Dan,
> > > > Re your comment on point 3):
> > > > The section you quote from MOF 1.3 is in the section about
> > > > the MOF Model
> > > > itself (the M3): it's in effect saying that although the
> > MOF M3 is an
> > > > instance of itself, it is (voluntarily) restricting itself to
> > > > only defining
> > > > attributes based on Data Types.
> > > > This is placing no such constraint on M2's such as CWM.
> > > > The (M3) class 'Attribute' inherits the reference 'type' from
> > > > TypedElement
> > > > which refers to Classifier. For Attribute the only
> > constraint on this
> > > > reference is at the TypedElement level where [C-13]
> > states that "An
> > > > Association may not be the type of a TypedElement"
> > > > (implicitly allowing the
> > > > other 2 subtypes of Classifier - Class and DataType).
> > > >
> > > > I'm not sure this addresses the original question since I was
> > > > not involved
> > > > in that discussion at Danvers: but CWM class 'Expression' may
> > > > legally be
> > > > used as the type of a CWM-defined attribute. Whether it's the
> > > > best way of
> > > > modeling the requirement is another question (e.g. it
> > does not allow a
> > > > (possibly complex or volatile) Expression (e.g. "Sales tax in
> > > > California")
> > > > to be defined only once and referred to from many places,
> > > > minimizing the
> > > > impact of any change).
> > > >
> > > > Pete
> > > >
> > > > Pete Rivett (pete.rivett@adaptive.com)
> > > > Chief Technology Officer, Adaptive Ltd
> > > > Dean Park House, 8-10 Dean Park Crescent, Bournemouth,
> BH1 1HL, UK
> > > > Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
> > > > http://www.adaptive.com
> > > >
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > > > > Sent: 19 July 2001 01:27
> > > > > To: Tolbert, Doug M
> > > > > Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> > > > > Subject: Re: "Baisley" questions from Danvers CWM RTF
meeting
> > > > >
> > > > >
> > > > >
> > > > > Doug,
> > > > >
> > > > > My comment is embedded below (==>).
> > > > >
> > > > > Regards,  Dan
> > > > >
> > > > > e-business Data Technology and Standard
> > > > > IBM Silicon Valley Laboratory
> > > > > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > > > > Internet:  dtchang@us.ibm.com
> > > > > VM:          IBMUSM50(DTCHANG)
> > > > > Phone:    (408)-463-2319
> > > > >
> > > > --cut--
> > > > >
> > > > >
> > > > > (3) Expressions as datatypes.  At first blush, Don doesn't
> > > > > think this is
> > > > > illegal (as was discussed in Danvers).  Consequently, I've
> > > > > posted an issue
> > > > > for the RTF on the OMG issues list about this topic so that
> > > > it can be
> > > > > considered more fully.  Don and I will take a more detailed
> > > > > look at it in
> > > > > the near future.
> > > > > ==> Please see the last paragraph on page 56 of MOF 1.3: All
> > > > > (Attributes)
> > > > > have a "type" that is
> > > > > ==> represented using a Data Type ..., and page 57 on
> > > > "type:": This is
> > > > > expressed as the name of
> > > > > ==> a Data Type defined in Section 3.6, or as the name of a
> > > > > CORBA data type
> > > > > ...
> > > > > ==> Therefore, it is illegal to use a CWM-defined Expression
> > > > > (an instance
> > > > > of a MOF Class) as the
> > > > > ==> type in defining a CWM attribute (an instance of a MOF
> > > > attribute).
> > > > > ==> I am copying Sridhar on this note so that he can confirm
> > > > > or rebuke what
> > > > > are stated in MOF 1.3.
> > > > >
> > > > > Doug
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > The information contained in this email and any
> attached files are
> > > > confidential and intended solely for the addressee(s). The
> > > > e-mail may be
> > > > legally privileged or prohibited from disclosure and
> > unauthorised use.
> > > > If you are not the named addressee you may not use, copy or
> > > > disclose this
> > > > information to any other person.  If you received this
> > > > message in error
> > > > please notify the sender immediately.
> > > > Any views or opinions presented here may be solely those of
> > > > the originator
> > > > and do not necessarily reflect those of the Company.
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > The information contained in this email and any attached files
are
> > > confidential and intended solely for the addressee(s). The
> > e-mail may be
> > > legally privileged or prohibited from disclosure and
> > unauthorised use.
> > > If you are not the named addressee you may not use, copy or
> > disclose this
> > > information to any other person.  If you received this
> > message in error
> > > please notify the sender immediately.
> > > Any views or opinions presented here may be solely those of the
> > originator
> > > and do not necessarily reflect those of the Company.
> > >
> > >
> > >
> > >
> >
> >
> >
> >
> >
>
>
>
>
>
>


The information contained in this email and any attached files are
confidential and intended solely for the addressee(s). The e-mail may
be legally privileged or prohibited from disclosure and unauthorised
use.
If you are not the named addressee you may not use, copy or disclose
this information to any other person.  If you received this message in
error please notify the sender immediately.
Any views or opinions presented here may be solely those of the
originator and do not necessarily reflect those of the Company.

From: "Martin Matula" <martin.matula@czech.sun.com>
To: "Pete Rivett" <pete.rivett@adaptive.com>,
   "'Dan Chang'" <dtchang@us.ibm.com>,
   "'Baisley, Donald E'" <Donald.Baisley@unisys.com>
Cc: <cwm-rtf@omg.org>
References: <003501c1114b$288811b0$114c04c8@CHIMAY>
Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
Date: Fri, 20 Jul 2001 21:00:05 +0200
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.2919.6700
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: i07e9+DH!!8`V!!:n^!!

Hi,

> Expression is indeed an instance of MOF Class not MOF DataType: the
  way
you
> can tell is that the 'real' MOF datatypes have the <<datatype>>
stereotype:
> Multiplicity and Expression do not. Also because Expression has lots
  of
> subtypes - which are not allowed for MOF Datatypes (MOF constraint
[C-19]).
> If you look at the UML 1.4 XMI file, Multiplicity and Expression are
present
> as instances of Model.Class, and in the IDL as Interfaces. I'll
  email you
> the files privately since sending files to OMG mailing lists is not
> permitted.
> You're right that this seems an unheralded change from UML 1.3.

AFAIK the UML 1.3 models it the same way (i.e. both Expression and
Multiplicity are instances of MOF Classes rather than instances of MOF
DataType also in UML 1.3).

Regards,
Martin

Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
To: "Pete Rivett" <pete.rivett@adaptive.com>
Cc: cwm-rtf@omg.org, "'Baisley, Donald E'" <Donald.Baisley@unisys.com>
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OF2FBBA695.0FB9C326-ON88256A8F.006AFD79@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Fri, 20 Jul 2001 12:40:01 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
|June 18, 2001) at
 07/20/2001 01:40:02 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: Wbd!!X`Be9?>&!!e;id9


Pete,

Thank you for sending me the UML 1.4 IDL files. Yes, they have made
"an
unheralded change from UML 1.3", at least in IDL mapping. So now,
Name,
LocationReference still map to string (as in UML 1.3), but
Expression/BooleanExpression/... map to interface (different from UML
1.3).
I am waiting to receive the UML 1.4 spec from you and see how they
indicate
the difference between Name and Expression. In my version, there is no
difference.

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                           
                    "Pete Rivett"                                                                          
                    <pete.rivett@ada       To:     Dan Chang/Santa
                    Teresa/IBM@IBMUS, "'Baisley, Donald E'" 
                    ptive.com>
                    <Donald.Baisley@unisys.com>                                    
                                           cc:     <cwm-rtf@omg.org>                                       
                    07/20/01 11:38         Subject:     RE: "Baisley"
                    questions from Danvers CWM RTF       
                    AM                      meeting                                                        
                                                                                                           
                                                                                                           
                                                                                                           



Dan,
Expression is in the UML "Datatype" package but "datatype" is here
just a
package name: it's not necessarily the case that the contents of this
package must all instances of MOF Datatype.

Expression is indeed an instance of MOF Class not MOF DataType: the
way you
can tell is that the 'real' MOF datatypes have the <<datatype>>
stereotype:
Multiplicity and Expression do not. Also because Expression has lots
of
subtypes - which are not allowed for MOF Datatypes (MOF constraint
[C-19]).
If you look at the UML 1.4 XMI file, Multiplicity and Expression are
present
as instances of Model.Class, and in the IDL as Interfaces. I'll email
you
the files privately since sending files to OMG mailing lists is not
permitted.
You're right that this seems an unheralded change from UML 1.3.

In another email you pointed out that:
"... in UML, which is M2, only data types defined in the Data Type
package
can be used to declare
the types of class attributes." I quickly looked in the UML 1.4 spec
but
could not find this restriction: I'd be interested to discover whether
this
is a Constraint in UML or a convention/policy that has been adopted?

Pete

> -----Original Message-----
> From: Dan Chang [mailto:dtchang@us.ibm.com]
> Sent: 20 July 2001 17:46
> To: Baisley, Donald E
> Cc: cwm-rtf@omg.org
> Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
>
>
>
> Don,
>
> Unfortunately, it does not.
>
> First, in the draft UML 1.4 spec dated 1/01 that I have,
> Expression, etc.
> are still in the Data Type package. Nothing says that they
> are different
> from UML 1.3. If you have a newer version that is different,
> please send it
> to me or the CWM RTF (I could not download the 2/01 version
> from the OMG
> web site)..
>
> Second, whether Expression ,etc. are data types (mapped to struct or
> equivalent) or classes (mapped to interface/class) has implication
on
> modeling and usage/performance. For example, the CWM ExpressionNode
> inherits from CWM Element and not CWM Expression. The
> assumption was that
> CWM Expression, as Expression in UML 1.3, cannot be specialized like
a
> normal class.
>
> Regards,  Dan
>
> e-business Data Technology and Standard
> IBM Silicon Valley Laboratory
> Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> Internet:  dtchang@us.ibm.com
> VM:          IBMUSM50(DTCHANG)
> Phone:    (408)-463-2319
>
>
>
>
>                     "Baisley, Donald
>
>                     E"                      To:
> cwm-rtf@omg.org
>                     <Donald.Baisley@u       cc:
>
>                     nisys.com>              Subject:     RE:
> "Baisley" questions from Danvers CWM RTF
>                                              meeting
>
>                     07/19/01 05:59 PM
>
>
>
>
>
>
>
>
> Hi,
>
> Since my name is in the subject of this email thread I
> thought I should
> interject something.  As Pete has pointed out, section 3.2 of the
MOF
> Specification does not prevent a class being the type of an
attribute.
> Section 3.2 merely describes how the MOF Model is described in the
> specification.  It just happens that MOF Model does not ever
> use a class as
> the type of an attribute, so the wording refers to data types.
>
> MOF explicitly allows an attribute's type to be a class.  The UML
1.4
> metamodel uses this capability with Expression and
> Multiplicity, both of
> which are classes (not data types) in UML 1.4.  According to
> MOF rules,
> there are composition semantics when an attribute's type is a
> class.  For
> example, if an expression is assigned as the default value of
> a parameter,
> then that expression belongs to the parameter and cannot be
> owned by any
> other object (just as if a composite association was used
> between parameter
> and expression).
>
> I hope this clears things up.
>
> Regards,
> Don
>
> -----Original Message-----
> From: Iyengar, Sridhar
> Sent: Thursday, July 19, 2001 12:03 PM
> To: Don Baisley; Doug Tolbert
> Subject: FW: "Baisley" questions from Danvers CWM RTF meeting
> Importance: High
>
>
> Can we calm down Dan - what is his problem?
>
> --------------------------------------------------------------
> ----------
> ----
> Sridhar Iyengar
> Unisys Fellow, Director of Advanced Technology
> 25725, Jeronimo Road
> Mission Viejo, CA 92691
>
> E-mail : Sridhar.iyengar2@unisys.com
> Phone : 949-380-5692
> Fax     : 949-380-6632
> --------------------------------------------------------------
> ----------
> --------------
>
>
>
>
> > -----Original Message-----
> > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > Sent: Thursday, July 19, 2001 11:58 AM
> > To: Martin Matula
> > Cc: cwm-rtf@omg.org; 'Tolbert, Doug M';
> John_Poole@hyperion.com; Pete
> > Rivett; 'Iyengar, Sridhar'
> > Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> >
> > Martin,
> >
> > If so, then we need to re-examine CWM.
> >
> > For example, the CWM Object Model is meant to be a subset of
> > UML. However,
> > in UML, Expression is a Data Type and, in the IDL binding, it
> > maps to a
> > struct. In CWM, Expression is not a Data Type and, in the IDL
> > binding, it
> > maps to an interface. Also, in UML, all Expresions (Expression,
> > BooleanExpression, ProcedureExpression, ...) have the same
> > "attributes"
> > (language and body), map to the identical struct, and cannot
> > be specialized
> > as normal classes. In CWM, Expressions (Expression,
> BooleanExpression,
> > ProcudureExpression) can have different attributes, map to
different
> > interfaces, and can be specialized as normal classes.
> >
> > We definitely did not intend to make such modeling changes
> > (on Expressions,
> > Multiplicity, MultiplicityRange, etc.) when we derived the
> CWM Object
> > Model.
> >
> > Regards,  Dan
> >
> > e-business Data Technology and Standard
> > IBM Silicon Valley Laboratory
> > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > Internet:  dtchang@us.ibm.com
> > VM:          IBMUSM50(DTCHANG)
> > Phone:    (408)-463-2319
> >
> >
> >
> >
> >                     "Martin Matula"
> >
> >                     <martin.matula@czec       To:     "Pete
> > Rivett" <pete.rivett@adaptive.com>, Dan
> >                     h.sun.com>                 Chang/Santa
> > Teresa/IBM@IBMUS
> >                                               cc:
> > <cwm-rtf@omg.org>, "'Tolbert, Doug M'"
> >                     07/19/01 10:58 AM
> > <Doug.Tolbert@unisys.com>, <John_Poole@hyperion.com>,
> >                                                "'Iyengar,
> > Sridhar'" <Sridhar.Iyengar2@unisys.com>
> >                                               Subject:
> > Re: "Baisley" questions from Danvers CWM RTF
> >                                                meeting
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > Hi Dan,
> > if CWM attribute means "M2 instance of MOF Attribute in your
> > CWM metamodel"
> > then Pete is correct. Any instance of MOF Class or MOF
> DataType can be
> > referenced as type from any instance of MOF Attribute in your
model.
> > This means, at M2 level you can model an attribute in CWM
> > metamodel to be
> > of
> > type CWM Expression. It will allow you to store instances of CWM
> > expressions
> > in this attribute at M1 level.
> > Martin
> >
> > ----- Original Message -----
> > From: "Dan Chang" <dtchang@us.ibm.com>
> > To: "Pete Rivett" <pete.rivett@adaptive.com>
> > Cc: <cwm-rtf@omg.org>; "'Tolbert, Doug M'"
> <Doug.Tolbert@unisys.com>;
> > <John_Poole@hyperion.com>; "'Iyengar, Sridhar'"
> > <Sridhar.Iyengar2@unisys.com>
> > Sent: Thursday, July 19, 2001 7:25 PM
> > Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
> >
> >
> > >
> > > Pete,
> > >
> > > I hope Sridhar can clarify what MOF 1.3 means.
> > >
> > > CWM Expression inherits from CWM Element and is not a CWM
> > Classifier.
> > > Therefore, it cannot be used as the type of a M1 CWM Attribute.
> > >
> > > If CWM Expression, being an instance of a MOF Class, can be
> > used as the
> > > type of a M2 CWM attribute (an instance of a MOF Attribute),
then
> > anything
> > > in CWM (being an instance of a MOF Class) can be used as
> > the type of a M2
> > > CWM attribute. For example, a CWM Element, a CWM Attribute, a
CWM
> > > Generalization, a CWM Association, a CWM AssociationEnd, etc.
> > >
> > > Are you sure?
> > >
> > > Regards,  Dan
> > >
> > > e-business Data Technology and Standard
> > > IBM Silicon Valley Laboratory
> > > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > > Internet:  dtchang@us.ibm.com
> > > VM:          IBMUSM50(DTCHANG)
> > > Phone:    (408)-463-2319
> > >
> > >
> > >
> > >                     "Pete Rivett"
> > >                     <pete.rivett@ada       To:
> > <John_Poole@hyperion.com>
> > >                     ptive.com>             cc:     Dan
Chang/Santa
> > Teresa/IBM@IBMUS, "'Tolbert, Doug M'"
> > >
> > <Doug.Tolbert@unisys.com>,
> > <cwm-rtf@omg.org>, "'Iyengar,
> > >                     07/19/01 08:20          Sridhar'"
> > <Sridhar.Iyengar2@unisys.com>
> > >                     AM                     Subject:     RE:
> > "Baisley"
> > questions from Danvers CWM RTF
> > >                                             meeting
> > >
> > >
> > >
> > >
> > >
> > >
> > > John,
> > > I hope I made the point in my previous mail that it is NOT
> > illegal to use
> > > CWM's Expression class as the type of an attribute.
> > >
> > > Nonetheless your conditional statement is correct -
> > although for clarity
> > > I'd
> > > replace "a CWM Expression" (which could be read as "an
> > instance of the
> > > class
> > > Expression") with "CWM's Expression class".
> > >
> > > Pete
> > >
> > > > -----Original Message-----
> > > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com]
> > > > Sent: 19 July 2001 15:48
> > > > To: Pete Rivett
> > > > Cc: 'Dan Chang'; 'Tolbert, Doug M'; cwm-rtf@omg.org;
> > > > 'Iyengar, Sridhar'
> > > > Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
> > > >
> > > >
> > > >
> > > >
> > > > Hi all,
> > > >
> > > > It seems to me, by way of analogy, that if it is illegal for
> > > > me to use a
> > > > CWM Expression as an attribute type in another CWM class,
> > > > then it would be
> > > > equally illegal for me to use any other CWM class as the
> > type of an
> > > > object-valued attribute in another CWM class (e.g.,
> > ExpressionNode).
> > > >
> > > > Is this a reasonable conclusion, or am I mistaken about this?
> > > >
> > > > - John
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > "Pete Rivett" <pete.rivett@adaptive.com> on 07/18/2001
> 10:49:57 PM
> > > >
> > > > To:   "'Dan Chang'" <dtchang@us.ibm.com>, "'Tolbert, Doug M'"
> > > >       <Doug.Tolbert@unisys.com>
> > > > cc:   <cwm-rtf@omg.org>, "'Iyengar, Sridhar'"
> > > > <Sridhar.Iyengar2@unisys.com>
> > > > Subject:  RE: "Baisley" questions from Danvers CWM RTF meeting
> > > >
> > > >
> > > > Dan,
> > > > Re your comment on point 3):
> > > > The section you quote from MOF 1.3 is in the section about
> > > > the MOF Model
> > > > itself (the M3): it's in effect saying that although the
> > MOF M3 is an
> > > > instance of itself, it is (voluntarily) restricting itself to
> > > > only defining
> > > > attributes based on Data Types.
> > > > This is placing no such constraint on M2's such as CWM.
> > > > The (M3) class 'Attribute' inherits the reference 'type' from
> > > > TypedElement
> > > > which refers to Classifier. For Attribute the only
> > constraint on this
> > > > reference is at the TypedElement level where [C-13]
> > states that "An
> > > > Association may not be the type of a TypedElement"
> > > > (implicitly allowing the
> > > > other 2 subtypes of Classifier - Class and DataType).
> > > >
> > > > I'm not sure this addresses the original question since I was
> > > > not involved
> > > > in that discussion at Danvers: but CWM class 'Expression' may
> > > > legally be
> > > > used as the type of a CWM-defined attribute. Whether it's the
> > > > best way of
> > > > modeling the requirement is another question (e.g. it
> > does not allow a
> > > > (possibly complex or volatile) Expression (e.g. "Sales tax in
> > > > California")
> > > > to be defined only once and referred to from many places,
> > > > minimizing the
> > > > impact of any change).
> > > >
> > > > Pete
> > > >
> > > > Pete Rivett (pete.rivett@adaptive.com)
> > > > Chief Technology Officer, Adaptive Ltd
> > > > Dean Park House, 8-10 Dean Park Crescent, Bournemouth,
> BH1 1HL, UK
> > > > Tel:  +44 (0)1202 449419 Fax: +44 (0)1202 449448
> > > > http://www.adaptive.com
> > > >
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: Dan Chang [mailto:dtchang@us.ibm.com]
> > > > > Sent: 19 July 2001 01:27
> > > > > To: Tolbert, Doug M
> > > > > Cc: cwm-rtf@omg.org; Iyengar, Sridhar
> > > > > Subject: Re: "Baisley" questions from Danvers CWM RTF
meeting
> > > > >
> > > > >
> > > > >
> > > > > Doug,
> > > > >
> > > > > My comment is embedded below (==>).
> > > > >
> > > > > Regards,  Dan
> > > > >
> > > > > e-business Data Technology and Standard
> > > > > IBM Silicon Valley Laboratory
> > > > > Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
> > > > > Internet:  dtchang@us.ibm.com
> > > > > VM:          IBMUSM50(DTCHANG)
> > > > > Phone:    (408)-463-2319
> > > > >
> > > > --cut--
> > > > >
> > > > >
> > > > > (3) Expressions as datatypes.  At first blush, Don doesn't
> > > > > think this is
> > > > > illegal (as was discussed in Danvers).  Consequently, I've
> > > > > posted an issue
> > > > > for the RTF on the OMG issues list about this topic so that
> > > > it can be
> > > > > considered more fully.  Don and I will take a more detailed
> > > > > look at it in
> > > > > the near future.
> > > > > ==> Please see the last paragraph on page 56 of MOF 1.3: All
> > > > > (Attributes)
> > > > > have a "type" that is
> > > > > ==> represented using a Data Type ..., and page 57 on
> > > > "type:": This is
> > > > > expressed as the name of
> > > > > ==> a Data Type defined in Section 3.6, or as the name of a
> > > > > CORBA data type
> > > > > ...
> > > > > ==> Therefore, it is illegal to use a CWM-defined Expression
> > > > > (an instance
> > > > > of a MOF Class) as the
> > > > > ==> type in defining a CWM attribute (an instance of a MOF
> > > > attribute).
> > > > > ==> I am copying Sridhar on this note so that he can confirm
> > > > > or rebuke what
> > > > > are stated in MOF 1.3.
> > > > >
> > > > > Doug
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > The information contained in this email and any
> attached files are
> > > > confidential and intended solely for the addressee(s). The
> > > > e-mail may be
> > > > legally privileged or prohibited from disclosure and
> > unauthorised use.
> > > > If you are not the named addressee you may not use, copy or
> > > > disclose this
> > > > information to any other person.  If you received this
> > > > message in error
> > > > please notify the sender immediately.
> > > > Any views or opinions presented here may be solely those of
> > > > the originator
> > > > and do not necessarily reflect those of the Company.
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > The information contained in this email and any attached files
are
> > > confidential and intended solely for the addressee(s). The
> > e-mail may be
> > > legally privileged or prohibited from disclosure and
> > unauthorised use.
> > > If you are not the named addressee you may not use, copy or
> > disclose this
> > > information to any other person.  If you received this
> > message in error
> > > please notify the sender immediately.
> > > Any views or opinions presented here may be solely those of the
> > originator
> > > and do not necessarily reflect those of the Company.
> > >
> > >
> > >
> > >
> >
> >
> >
> >
> >
>
>
>
>
>
>


The information contained in this email and any attached files are
confidential and intended solely for the addressee(s). The e-mail may
be
legally privileged or prohibited from disclosure and unauthorised use.
If you are not the named addressee you may not use, copy or disclose
this
information to any other person.  If you received this message in
error
please notify the sender immediately.
Any views or opinions presented here may be solely those of the
originator
and do not necessarily reflect those of the Company.

Subject: Re: "Baisley" questions from Danvers CWM RTF meeting
To: "Martin Matula" <martin.matula@czech.sun.com>
Cc: cwm-rtf@omg.org, "'Baisley, Donald E'"
<Donald.Baisley@unisys.com>,
   "Pete Rivett" <pete.rivett@adaptive.com>
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OFC7755C49.A01589F7-ON88256A8F.006BCBD8@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Fri, 20 Jul 2001 12:42:15 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
|June 18, 2001) at
 07/20/2001 01:42:15 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: E[gd90V!!!LYb!!'[)!!


Martin,

Not so in UML 1.3. There, Expression, Multiplicity, etc. are not
normal
classes and do not mapped to interfaces in IDL. They map to structs.

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                              
                    "Martin Matula"                                                                           
                    <martin.matula@czec       To:     "Pete Rivett"
                    <pete.rivett@adaptive.com>, Dan           
                    h.sun.com>                 Chang/Santa
                    Teresa/IBM@IBMUS, "'Baisley, Donald E'"            
                                               <Donald.Baisley@unisys.com>                                    
                    07/20/01 12:00 PM         cc:
                    <cwm-rtf@omg.org>                                       
                                              Subject:     Re:
                    "Baisley" questions from Danvers CWM RTF       
                                               meeting                                                        
                                                                                                              
                                                                                                              
                                                                                                              



Hi,

> Expression is indeed an instance of MOF Class not MOF DataType: the
  way
you
> can tell is that the 'real' MOF datatypes have the <<datatype>>
stereotype:
> Multiplicity and Expression do not. Also because Expression has lots
  of
> subtypes - which are not allowed for MOF Datatypes (MOF constraint
[C-19]).
> If you look at the UML 1.4 XMI file, Multiplicity and Expression are
present
> as instances of Model.Class, and in the IDL as Interfaces. I'll
  email you
> the files privately since sending files to OMG mailing lists is not
> permitted.
> You're right that this seems an unheralded change from UML 1.3.

AFAIK the UML 1.3 models it the same way (i.e. both Expression and
Multiplicity are instances of MOF Classes rather than instances of MOF
DataType also in UML 1.3).

Regards,
Martin

From: "Baisley, Donald E" <Donald.Baisley@unisys.com>
To: Dan Chang <dtchang@us.ibm.com>
Cc: cwm-rtf@omg.org
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Date: Fri, 20 Jul 2001 15:49:57 -0500
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2653.19)
Content-Type: text/plain;
	      charset="iso-8859-1"
X-UIDL: Q2E!!5[fd9I*$e9j2T!!

Hi Dan,

UML 1.3 used struct types in some places, such as for Expression.
This is
why we could not use UML 1.3 Expression as a supertype of CWM classes
back
when we were trying to build on UML.  In UML 1.4 all struct types were
replaced with classes, but the package structure was not changed (the
UML
1.4 RTF had decided to postpone all repacking issues).  The UML 1.4
specification includes a normative XML representation (ad/01-02-15)
which
clearly shows Expression to be a class, not a struct.  The generated
IDL
(ad/01-02-17) and DTD (ad/01-02-16) also show this to be the case.

The CWM metamodel's Expression is also a class.  This is clearly seen
in the
normative CWM XML (ad/01-02-03):

      <Model:Class xmi.id="a51" name="Expression" ...

Expression in UML 1.4 or in CWM 1.0 is a class and can be specialized
as a
class.

Best regards,
Don

-----Original Message-----
From: Dan Chang [mailto:dtchang@us.ibm.com]
Sent: Friday, July 20, 2001 9:46 AM
To: Baisley, Donald E
Cc: cwm-rtf@omg.org
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Importance: High



Don,

Unfortunately, it does not.

First, in the draft UML 1.4 spec dated 1/01 that I have, Expression,
etc.
are still in the Data Type package. Nothing says that they are
different
from UML 1.3. If you have a newer version that is different, please
send it
to me or the CWM RTF (I could not download the 2/01 version from the
OMG
web site)..

Second, whether Expression ,etc. are data types (mapped to struct or
equivalent) or classes (mapped to interface/class) has implication on
modeling and usage/performance. For example, the CWM ExpressionNode
inherits from CWM Element and not CWM Expression. The assumption was
that
CWM Expression, as Expression in UML 1.3, cannot be specialized like a
normal class.

Regards,  Dan


Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
To: "Baisley, Donald E" <Donald.Baisley@unisys.com>
Cc: cwm-rtf@omg.org
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OFF10C728B.58CB01C1-ON88256A8F.007C07FF@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Fri, 20 Jul 2001 15:53:18 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
|June 18, 2001) at
 07/20/2001 04:53:19 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: EDIe9'KZ!!!4?e9eJHe9


Don,

Unfortunately this major change in design in UML, though clearly
represented in the XML file, is not clearly specified in the
specification
or conveyed to the CWM team. Therefore, we have the confusion that we
have
had and the situation in CWM 1.0 that Expression is a class but
ExpressionNode still does not inherit from Expression.

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                            
                    "Baisley, Donald                                                                        
                    E"                      To:     Dan Chang/Santa
                    Teresa/IBM@IBMUS                        
                    <Donald.Baisley@u       cc:     cwm-rtf@omg.org                                         
                    nisys.com>              Subject:     RE: "Baisley"
                    questions from Danvers CWM RTF       
                                             meeting                                                        
                    07/20/01 01:49 PM                                                                       
                                                                                                            
                                                                                                            



Hi Dan,

UML 1.3 used struct types in some places, such as for Expression.
This is
why we could not use UML 1.3 Expression as a supertype of CWM classes
back
when we were trying to build on UML.  In UML 1.4 all struct types were
replaced with classes, but the package structure was not changed (the
UML
1.4 RTF had decided to postpone all repacking issues).  The UML 1.4
specification includes a normative XML representation (ad/01-02-15)
which
clearly shows Expression to be a class, not a struct.  The generated
IDL
(ad/01-02-17) and DTD (ad/01-02-16) also show this to be the case.

The CWM metamodel's Expression is also a class.  This is clearly seen
in
the
normative CWM XML (ad/01-02-03):

      <Model:Class xmi.id="a51" name="Expression" ...

Expression in UML 1.4 or in CWM 1.0 is a class and can be specialized
as a
class.

Best regards,
Don

-----Original Message-----
From: Dan Chang [mailto:dtchang@us.ibm.com]
Sent: Friday, July 20, 2001 9:46 AM
To: Baisley, Donald E
Cc: cwm-rtf@omg.org
Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
Importance: High



Don,

Unfortunately, it does not.

First, in the draft UML 1.4 spec dated 1/01 that I have, Expression,
etc.
are still in the Data Type package. Nothing says that they are
different
from UML 1.3. If you have a newer version that is different, please
send it
to me or the CWM RTF (I could not download the 2/01 version from the
OMG
web site)..

Second, whether Expression ,etc. are data types (mapped to struct or
equivalent) or classes (mapped to interface/class) has implication on
modeling and usage/performance. For example, the CWM ExpressionNode
inherits from CWM Element and not CWM Expression. The assumption was
that
CWM Expression, as Expression in UML 1.3, cannot be specialized like a
normal class.

Regards,  Dan

Subject: RE: "Baisley" questions from Danvers CWM RTF meeting
To: "Iyengar, Sridhar" <Sridhar.Iyengar2@unisys.com>
Cc: "Stephen Brodsky" <sbrodsky@us.ibm.com>, cwm-rtf@omg.org,
   "Tolbert, Doug M" <Doug.Tolbert@unisys.com>
X-Mailer: Lotus Notes Release 5.0.7  March 21, 2001
Message-ID: <OFAD076940.9628BFF0-ON88256A8F.008106FF@boulder.ibm.com>
From: "Dan Chang" <dtchang@us.ibm.com>
Date: Fri, 20 Jul 2001 16:35:26 -0700
X-MIMETrack: Serialize by Router on D03NM069/03/M/IBM(Release 5.0.8
|June 18, 2001) at
 07/20/2001 05:35:31 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-UIDL: K(?!!a~Pe97MCe9RnHe9


Sridhar,

I talked to Steve. He has agreed to provide the generation of XML, DTD
and
XSD for CWM, starting with CWM 1.1. I will continue to rely on your
team
(Doug) to generate IDL for CWM. Thanks in advance.

Regards,  Dan

e-business Data Technology and Standard
IBM Silicon Valley Laboratory
Notes:     Dan Chang/Santa Teresa/IBM@IBMUS
Internet:  dtchang@us.ibm.com
VM:          IBMUSM50(DTCHANG)
Phone:    (408)-463-2319


                                                                                                              
                    "Iyengar, Sridhar"                                                                        
                    <Sridhar.Iyengar2@u       To:     Dan Chang/Santa
                    Teresa/IBM@IBMUS, "Tolbert, Doug M"     
                    nisys.com>
                    <Doug.Tolbert@unisys.com>                                      
                                              cc:     cwm-rtf@omg.org,
                    "Iyengar, Sridhar"                     
                    07/18/01 06:49 PM
                    <Sridhar.Iyengar2@unisys.com>                                  
                                              Subject:     RE:
                    "Baisley" questions from Danvers CWM RTF       
                                               meeting                                                        
                                                                                                              
                                                                                                              
                                                                                                              



Regarding Dan's question on Unisys help on IDL/XML generation - Here
is the
Unisys position.

Based on resources available we (Unisys) will assist in some of these
generations. But several vendors including IBM, Adaptive, SUN and
Oracle
have or will soon have this technology. So hopefully the xsd
generation is
a
short term issue.

P.S.  The ADTF has not yet voted on XMI2.  This will happen in
Toronto.

Sridhar



------------------------------------------------------------------------
----
Sridhar Iyengar
Unisys Fellow, Director of Advanced Technology
25725, Jeronimo Road
Mission Viejo, CA 92691

E-mail : Sridhar.iyengar2@unisys.com
Phone : 949-380-5692
Fax     : 949-380-6632
------------------------------------------------------------------------
-
>