E2 to E1 SHTOLO test files

Tue May 7 18:38:32 EDT 2002

All,

    Just to jump in again, I think the situation we have is a bit more
complex than Phil likes to think of it.

    I know that I originally mentioned files, but that is not the basic
point. We have to deal with a collection of schemas, whether they are in one
file or multiple files.

    STEP has a collection of schemas, most (all?) of which are Ed1. Modules
are planning on using Ed2, and I am also producing Ed1 IRs
leading to an Ed2 AP (hopefully using modules).

    We can parse each schema seperately, but unless the schema is completely
self-contained there will be unidentified references. Generally speaking, it
will be necessary at some point to handle multiple schemas with no guarantee
that they will all be Ed1 or all Ed2; more likely in future the guarantee is
that they will be mix of Ed1 and Ed2.

    We do have to check that a schema corresponds to Ed1 or Ed2 as
appropriate, and that a mix of Ed1 and Ed2 schemas do form a complete model.
We either have to parse and `link' and link a collection
of Ed1 and Ed2 schemas, or we have to seperately parse a collection of Ed1
schemas and a collection of Ed2 schemas and then link the two collections.

    Any future Shtolo process will have to handle Ed1 and Ed2
simultaneously. 

    We must not force parser vendors into a parse and then link process, the
LRM must let them be free to decide how they can best deal with an Ed1/Ed2
collection.

    Granted Ed1 schemas can be parsed by an Ed2 parser, but the reverse is
not true. A parser must know what it is expected to parse. If Ed3 does
miraculously appear then the situation will be worse than it is now, as
since it plans on getting rid of SET it means that an Ed3 parser will not be
able to parse Ed1 or Ed2.

    Better fix the problem now. The EXPRESS Committee has had enough bitter
experiences in not fixing something that was bad.

    Ed suggested that either the language_version_id goes with "all the
schemas you need" or "each schema seperately". My answer is that it has to
be "each schema seperately".

Peter W.

Dr Peter R. Wilson
Boeing Commercial Airplanes
PO Box 3707, MS 6H-AF, Seattle, WA 98124-2207
(Package Delivery: MS 6H-AF, 1601 E. Valley Frontage Road, Renton, WA 98055)
Tel: (425) 237-3506, Fax: (425) 237-3428
Email: peter.r.wilson at boeing.com
--------------------------------
Any opinions expressed above are personal;
they shall not be construed as representative of any organisation.
--------------------------------
 

> -----Original Message-----
> From: Phil Spiby [mailto:Phil.Spiby at eurostep.com]
> Sent: Tuesday, May 07, 2002 2:17 PM
> To: Jochen Haenisch; 'Wilson, Peter R'
> Cc: Hans Karsten Dahl; wg11 at steptools.com; Hendrix, Thomas E; Arne Tøn
> Subject: RE: E2 to E1 SHTOLO test files
> 
> 
> Folks,
> 
> I remember the discussion with Peter etc. but my 
> understanding was that
> there was no change agreed to the DAM.
> 
> My personal understanding of this situation, which is open to 
> dispute, is
> the following:
> EXPRESS has deliberately avoided the use of the term file, since to
> acknowledge files, we then need a means of defining 
> relationships between
> files, like #include etc.
> The concept of file is probably wrong if you are considering 
> an interactive
> development environment (this was an aims of some of the 
> original EXPRESS
> committee).
> The way we tried to get around this was to introduce the idea 
> that some
> mechanism would enable either an environment or a collection 
> of files to be
> dealt with as a complete unit, and that this complete unit 
> would obey the
> rules of the Syntax and allow a scope for schema names to 
> exist within.
> Although much of these concepts were removed in the run-up to
> standardisation in 1994 we still have some legacies in the 
> standard with
> keywords such as MODEL and END_MODEL, and the Note under 10.3.13.
> 
> It is my belief that this schema aggregation should all obey the same
> language syntax, and this is why I think the syntax conforms 
> to a single
> language version. If we consider the now almost dead Edition3 
> then Peter's
> proposal would allow some schemas to correspond to Edition 1, some to
> edition2 and perhaps some to edition3, i.e. different schemas 
> would have to
> be conformance checked against different standards ISO 
> 10303-11 for editions
> 1 and 2 and ISO 20303 for edition 3!.
> 
> I understand that Peter wants this so that he can easily concatenate a
> number of files together then parse the result, and I have 
> some sympathy for
> this approach. But I feel that if any one schema is to be 
> parsed against
> edition two then the whole collection should be parsed at that level.
> Edition two is, by design and SC4 ruling, upwardly compatible 
> with edition
> one!
> 
> Phil
> 
> > -----Original Message-----
> > From: owner-wg11 at steptools.com [mailto:owner-wg11 at steptools.com]On
> > Behalf Of Jochen Haenisch
> > Sent: Tuesday, May 07, 2002 5:24 PM
> > To: 'Wilson, Peter R'
> > Cc: Hans Karsten Dahl; 'wg11 at steptools.com'; Hendrix, 
> Thomas E; Arne Tøn
> > Subject: RE: E2 to E1 SHTOLO test files
> >
> >
> > Peter,
> >
> > your proposal may well make sense! I just referred to what 
> I think is
> > specified in the DAM so far. It will probably be quite 
> common to generate
> > longforms out of mixes of e1 and e2 shortforms.
> >
> > However, do you think that it is necessary to include 
> schemas according to
> > different Express versions in the same file? Should it not 
> be enough to
> > specify the language identifier once only in the top of a file?
> >
> > Agreed, the LRM does not relate the production term 
> "syntax" to files. We
> > should probably be more explicit in e2?!? There seems to be 
> a source for
> > misunderstandings.
> >
> > I do not think that there is anything related in the issues 
> resolutions.
> >
> > Best 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:	Wilson, Peter R [SMTP:peter.r.wilson at boeing.com]
> > > Sent:	7. mai 2002 17:55
> > > To:	'Jochen Haenisch'
> > > Cc:	'wg11 at steptools.com'; Hendrix, Thomas E; Wilson, Peter R
> > > Subject:	RE: E2 to E1 SHTOLO test files
> > >
> > > Jochen,
> > >
> > >     I had a discussion at Myrtle Beach with Phil Spiby on the
> > question of
> > > putting the language_version_id before each schema. As far as I
> > remember,
> > > Phil said that the assumption was that the entire EXPRESS 
> model would be
> > > in
> > > a single file and that the language_version_id applied to 
> the file. In
> > > other
> > > words, the syntax production applies to a file. The IR53 
> model is in
> > > several
> > > files, and hence the multiple occurrences of the
> > language_version_id. The
> > > LRM makes no mention of file(s) containing EXPRESS code.
> > >
> > >     What happens if a model is partly Edition 1 (e.g., 
> the traditional
> > > IRs)
> > > and partly Edition 2 (e.g., new IRs like 52 and 53)?
> > >
> > >     I think that a language_version_id should be 
> associable with an
> > > individual schema, not just with this unexplained concept 
> of a file. If
> > > you
> > > like, change the syntax production to read:
> > > syntax = [ language_version_id ] schema_decl { 
> [language_version_id ]
> > > schema_decl } .
> > >
> > > Regards
> > > Peter W.
> > >
> > > PS. I have not yet had a chance to look at Phil's latest list
> > of issues to
> > > see if this is in there or not.
> > >
>