[Ccsds-omg-liaison] Fwd: RE: AB Evaluation of space-05/05/01

[Adrian J. Hooke]

>Date: Mon, 20 Jun 2005 10:51:20 +0200
>From: Desfray <Philippe.Desfray at softeam.fr>
>Subject: RE: AB Evaluation of space-05/05/01
>To: "'Chonoles, Michael J'" <michael.j.chonoles at lmco.com>,
>  conrad.bock at nist.gov, 'Sridhar Iyengar' <siyengar at us.ibm.com>
>Cc: ab at omg.org, dtc at omg.org,
>  'Simon C Gerry Contr Det 12/VOC' <Gerry.Simon at schriever.af.mil>,
>  pm-sig at omg.org, space at omg.org, ptc at omg.org
>
>Dear all,
>
>I share the high concern with Conrad and Michael:
> >>> Because this is a domain spec, I felt insisting/requiring reuse of
>significant parts of UML2 -
>may introduce terminology(and other metamodel dependencies) not necessarily
>familiar to experts in the space domain.
>
>This is in opposition with the overall idea of standardization. Domains do
>not live in isolation, and fragmenting the vocabulary/model specificities
>per domain apart from the required pieces adds unecessary barriers between
>these communities.
>
>Anyway, we always observe that in any domain there is a variety of
>vocabulary, and sometime standards. At least, OMG should stay consistent
>with himself by reusing common core vocabulary.
>
>====================================
>Philippe Desfray
>VP for R&D - SOFTEAM
>Tel:  (33) 01 53968400
>Fax: (33) 01 53968401
>144 Av. des champs Elysees 75008 PARIS
>www.softeam.com
>www.objecteering.com
>
>
>-----Message d'origine-----
>De : Chonoles, Michael J [mailto:michael.j.chonoles at lmco.com]
>Envoye : vendredi 17 juin 2005 21:13
>A : conrad.bock at nist.gov; Sridhar Iyengar
>Cc : ab at omg.org; dtc at omg.org; Simon C Gerry Contr Det 12/VOC;
>pm-sig at omg.org; space at omg.org; ptc at omg.org
>Objet : RE: AB Evaluation of space-05/05/01
>
>
>
>Especially with the development of SysML, it appears that much of the
>space community will be familiar with the UML2 or (slightly different)
>SysML modeling approaches in these areas.
>
>I've been very concerned with the potential of fragmenting the modeling
>communities into mutually-incomprehensible dialects. The point of
>standards is to bring down barriers by leveraging well-understood prior
>work. If we allow for different approaches (for no "good" reason) we'll
>be back to the world before UML,
>
>*       Where there were over 50 incompatible notations,
>*       Where companies and vendors will wary of investing in tools or
>         training because of the uncertainly of market outcome,
>*       Where lessons learned in one project could not be applied
>anywhere
>         else, and
>*       Where further research and progress were stymied by a lack of
>common  vocabulary and experience
>
>Michael Jesse Chonoles
>Coauthor UML 2 For Dummies
>Lockheed Martin MS2/IS&S
>
>-----Original Message-----
>From: Conrad Bock [mailto:conrad.bock at nist.gov]
>Sent: Friday, June 17, 2005 11:59 AM
>To: Sridhar Iyengar
>Cc: ab at omg.org; dtc at omg.org; Simon C Gerry Contr Det 12/VOC;
>pm-sig at omg.org; space at omg.org; ptc at omg.org
>Subject: RE: AB Evaluation of space-05/05/01
>
>Hi Sridhar,
>
>  > I agree that the concepts overlap with what is in UML2 (as well as
>  > common scripting/programming languages).  Because this is a domain
>  > spec, I felt insisting/requiring reuse of significant parts of UML2 -
>  > may introduce terminology(and other metamodel dependencies) not
>  > necessarily familiar to experts in the space domain.
>
>By that argument, none of the domain-independent models adopted in the
>PTF would be reused by the domains.
>
>Terminolgy is not much of a barrier, since space operations and the UML
>structured activities come from the same experience-base, ie, code.
>They both use the terms conditional, loops, expressions, pre and
>postconditions, parameters, control flow, and other language.
>
>In the few cases where there are terminology differences, profiles are
>available for renaming, or subtyping using MOF.  If OMG builds a new
>model for every domain that uses different terms for the same thing,
>we'll have quite a lot of non-interoperable models for the same
>concepts.  In the case of space operations, we'll have two models using
>the same names for the same things.
>
>Conrad
>
>
>-----Original Message-----
>From: Conrad Bock [mailto:conrad.bock at nist.gov]
>Sent: Friday, June 17, 2005 10:36 AM
>To: Sridhar Iyengar; ab at omg.org
>Cc: space at omg.org; dtc at omg.org; pm-sig at omg.org;
>Subject: RE: AB Evaluation of space-05/05/01
>
>Hi all,
>
>The RFP has only partially responded to my comments of March 22nd and
>April 7th, see below.  Requirement 6.5.11 has significant overlap with
>UML 1.5 and UML 2.  The RFP only generally mentions consideration of the
>UML models that fulfills the requirements.
>
>Conrad
>
>   A process/coding model is available in UML 1.5 and UML 2 that covers
>   this requirement:
>
>     6.5.11 The proposed metamodel shall allow for the definition of:
>       a) Control flow constructs
>       b) Expressions
>       c) Sub processes
>       d) Data elements and parameters
>       e) Pre and post conditions (including time)
>
>   The RFP currently only mentions "considering" UML, and does not ask
>   for an extension to it, MOF or profile.  Other than the specs
>   (doc.omg.org/formal/03-03-01, doc.omg.org/ptc/04-10-02), more
>   explanation of the UML 2 models covering the topics above is at:
>
>     http://www.jot.fm/issues/issue_2005_05/column4
>
>   and more at www.conradbock.org/#UML2.0.
>