FW: error resolution assistance needed
Phil Spiby
Phil.Spiby at eurostep.com
Tue Jan 21 06:25:40 EST 2003
Jochen,
You are correct.
We originally had under 9.2.6 that a type could be considered as a
specialisation of itself. This was removed in the last stages of balloting,
and we thought that had altered everywhere this affected such as assignment
compatibility etc.. It seems that we didn't catch all places since it was
assumed that attribute redeclaration allowed you to just remove the
optional, or convert an explicit into a derived WITHOUT modifying the type
of the attribute (and I believe the explicit to derived is done in a couple
of places in STEP IR's at present).
I think Ed is also correct in that the syntax is a little too verbose at
this stage for just renaming, but on the other hand when you get deep into
the modules, it is perhaps a good thing to have the full type information
along with the renaming, just so you don't have to follow through an
horrendous search to find out the type of the renamed attribute.
Just to re-iterate Jochen's question, if there are any protests against this
then please shout up now!
Phil
> -----Original Message-----
> From: Jochen Haenisch [mailto:Jochen.Haenisch at epmtech.jotne.com]
> Sent: 21 January 2003 10:07
> To: 'Phil.Spiby at Eurostep.com'; 'Hendrix, Thomas E'
> Cc: 'Tim King'; 'Ed'; Arne Tøn; 'WG11 exploder'
> Subject: RE: FW: error resolution assistance needed
>
>
> Phil,
>
> if this is the agreement in WG11, fine with me. Then we need
> to change (or at least: clarify) rule a) in 9.2.3.4 which
> currently says:
> "a) The attribute redeclared in the subtype shall be a
> specialization of the attribute of the same name in the supertype."
>
> Nontheless, 9.2.3.4 needs some additional text as pointed out
> in FR-8 explaining the renaming mechanism.
>
> If there are no protests against this view, I will edit P11e2
> accordingly.
>
> Regards, Jochen.
> ____________________________________________________________
> Jochen Haenisch E-mail:
> Jochen.Haenisch at epmtech.jotne.com
> EPM Technology AS Home of the EXPRESS Data Manager
> P.O Box 6629 Etterstad Tel: Int + 47 23 17 17
> 26; mobile: Int + 47
> 922 60 274
> N-0607 Oslo Fax: Int + 47 23 17 17 01
> Norway Web:
> http://www.epmtech.jotne.com
>
> > -----Original Message-----
> > From: Phil Spiby [SMTP:Phil.Spiby at Eurostep.com]
> > Sent: 21. januar 2003 11:01
> > To: 'Jochen Haenisch'; 'Hendrix, Thomas E'
> > Cc: 'Tim King'; 'Ed'; 'Arne Tøn'
> > Subject: RE: FW: error resolution assistance needed
> >
> > Jochen,
> >
> > If renaming without specialization is not in the document then it
> > should be! The re-declaration mechanism is to allow
> > - making optional attributes mandatory
> > - making explicit attributes derived
> > - constraining the type to a specialisation
> > And now in Ed2
> > - to rename an attribute
> >
> > There is, and should be, no statements that redeclaration
> is forced to
> > include specialisation!
> >
> > Phil
> >
> > PS I also haven't looked at the EXPRESS or the rest of this thread
> > yet, but wanted to state my
> > Understanding on what is written.
> >
> > > -----Original Message-----
> > > From: Jochen Haenisch [mailto:Jochen.Haenisch at epmtech.jotne.com]
> > > Sent: 20 January 2003 19:00
> > > To: 'Hendrix, Thomas E'
> > > Cc: 'Tim King'; 'Phil Spiby'; 'Ed'; Arne Tøn
> > > Subject: RE: FW: error resolution assistance needed
> > >
> > >
> > > Hei Tom,
> > >
> > > we can probably get into a lengthy discussion with this. So,
> > > Tim may rightaway choose a less controversial way of
> > > representing his requirement ... if the completion of the
> > > schema has some time constraints on it.
> > >
> > > I do not agree that P11e2 allows renaming without
> > > redeclaration. And redeclaration requires a specialization.
> > > The entire clause 9.2.3.4 is about redeclaration; and RENAMED
> > > appears only in this clause. I have no idea why there is no
> > > simple renaming, but as WG11 decisions currently stand, I can
> > > not see that this is possible.
> > >
> > > The agreed ballot comment resolutions for P11e2 are attached.
> > > Most of them are yet not incorporated into the final document
> > > ... . FR-8, which was rejected, relates probably to this
> > > case. Have a look.
> > >
> > > Best regards, Jochen.
> > > P.S.: I have not checked your Express - I anticipate that
> > > there is no specialization involved.
> > >
> > > <<ISO_10303-11_1994_DAM_all_comments.doc>>
> > > ____________________________________________________________
> > > Jochen Haenisch E-mail:
> > > Jochen.Haenisch at epmtech.jotne.com
> > > EPM Technology AS Home of the EXPRESS Data Manager
> > > P.O Box 6629 Etterstad Tel: Int + 47 23 17 17
> > > 26; mobile: Int + 47
> > > 922 60 274
> > > N-0607 Oslo Fax: Int + 47
> 23 17 17 01
> > > Norway Web:
> > > http://www.epmtech.jotne.com
> > >
> > > > -----Original Message-----
> > > > From: Hendrix, Thomas E [SMTP:thomas.e.hendrix at boeing.com]
> > > > Sent: 20. januar 2003 19:42
> > > > To: Jochen Haenisch (E-mail)
> > > > Subject: FW: FW: error resolution assistance needed
> > > >
> > > > Hei Jochen
> > > >
> > > > Tim King is getting this error from one of the PLCS ARM
> > > compilations.
> > > >
> > > > ERROR : Line 762: Illegal attribute redeclaration.
> > > > Redeclaring attribute:
> CONDITION_ASSERTION.VALUE_CONTEXT
> > > > RENAMED OBJECT_EVALUATED in line: 762
> > > > Redeclared attribute: VALUE_RECORD.VALUE_CONTEXT
> > > in line: 1064
> > > > Redeclaration is no specialisation, only renaming
> > > >
> > > > from this EXPRESS
> > > >
> > > > 757 ENTITY Condition_assertion
> > > > 758 SUBTYPE OF (Evaluation_record);
> > > > 759 SELF\Value_record.value_recorded RENAMED assertion :
> > > > Logical_assertion;
> > > > 760 SELF\Value_record.expression_evaluated RENAMED
> condition :
> > > > Condition;
> > > > 761 SELF\Value_record.value_context RENAMED
> object_evaluated :
> > > > OPTIONAL Entity_instance_reference;
> > > > 762 END_ENTITY;
> > > >
> > > > (see attachment for entire report)
> > > >
> > > > After reading Ed Barkeyers's reply to my note to wg11
> > > (below), I think
> > > > I will claim that the compiler should permit this. What do
> > > you think?
> > > >
> > > >
> > > > -----Original Message-----
> > > > From: Ed Barkmeyer [mailto:edbark at nist.gov]
> > > > Sent: Wednesday, January 15, 2003 9:52 AM
> > > > To: Hendrix, Thomas E
> > > > Cc: SC4 WG11
> > > > Subject: Re: FW: error resolution assistance needed
> > > >
> > > >
> > > > Thomas Hendrix wrote:
> > > > > 10303-11:1994 and its DAM 1 , 9.2.3.4
> > > > >
> > > > > "Rules and restrictions: a)" seems to disallow the use of
> > > > > RENAME
> > > > > for attribute redeclaration unless the renamed
> attribute is >
> > > > constrained by an additional change. Is this the
> intent? > If so,
> > > > propose relaxing this restriction.
> > > >
> > > > This is really a subtle point: rule (a) requires a
> > > "redeclaration" in
> > > > the full syntax thereof, but it is nowhere stated that a
> > > "redeclaration"
> > > > must actually constrain the data type of the attribute!
> > > >
> > > > The purpose of redeclaration (before, and therefore other than,
> > > > RENAME)
> > > > was to constrain the type, and there was formerly no reason
> > > to have a
> > > > redeclaration that does not further constrain the type. But a
> > > > redeclaration of an attribute that declares it to be of
> exactly the
> > > > original type appears to meet all the requirements for a valid
> > > > redeclaration.
> > > >
> > > > OTOH, it does seem that a simpler syntax for changing only the
> > > > name would be useful.
> > > >
> > > > -Ed
> > > >
> > > > --
> > > > Edward J. Barkmeyer Email: edbark at nist.gov
> > > > National Institute of Standards & Technology
> Manufacturing Systems
> > > > Integration Division
> > > > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3528
> > > > Gaithersburg, MD 20899-8260 FAX: +1 301-975-4694
> > > >
> > > > "The opinions expressed above do not reflect consensus of NIST,
> > > > and
> > > > have not been reviewed by any Government authority." << File:
> > > > arm_2003118v1_errors.zip >>
> > >
>
>
More information about the wg11
mailing list