Mail Archive Home | architecture List | May 2004 Index
| <-- Date Index --> | <-- Thread Index --> |
Hi,
The CORBA Component Model deals with heterogenous
components e.g. implemented in different programming
languages like C++, Java, etc. In this model, deployment
units are called Component and Assembly Archives, CAR and
AAR files respectively. These files are ZIP files.
A CAR is the deployment unit for only one component type
but could contain several implementations with several
associated meta data, eg. one implementation in Java and
one implementation in C++ for Windows and one implementation
on C++ for Linux X.Y.Z, etc.
An AAR is the deployment unit of a set of assembled components
and could refer to or embed several CAR files.
This is just to propose that 'Fractal deployment units'
could be called 'Fractal archives' independently of these
contents. Then we could speak about:
* instances : components seen as runtime units
* archives : components seen as deployment units
* implementations : components seen as development units
* types : components seen as design units
This could avoid us to reuse fully words like 'seed', 'germ',
or 'bundle'.
A+
Philippe Merle
Humberto Cervantes wrote:
>
> On Fri, 2004-05-07 at 17:53, Thierry COUPAYE wrote:
> > Le 7 mai 04, à 13:56, Sébastien Jean a écrit :
> >
> > >
> > >
> > >> Eric Bruneton wrote:
> > >>
> > >>> it is true that "packages", in this document, have nothing to do with
> > >>> Java packages. But "deployment unit" is a bit too long (I would
> > >>> prefer a single word). Any other idea?
> > >>>
> > >
> > > i propose either "seed" or "germ" that shortly stands for deployement
> > > unit
> >
> > "seed" or "germ" should be reserved for bootstrapping issues wrt a
> > deployment infrastructure IMHO. Maybe we can find something better than
> > 'package' but frankly, package is just fine for me. We would just have
> > to say 'java package' when speaking about java package (!) and package
> > otherwise. Sun takes words like 'interface' or 'package' and gives them
> > very strange semantics, i don't see why this should prevent us for
> > using these words which have a natural meaning to most people.
> >
> > T.
> >
> >
>
> Why don't you use the term "bundle"? I don't think it's patented by OSGi
> and the definition I found for it is pretty appropriate:
>
> "an assemblage of things that have been tied, bound, or packaged
> together."
>
> Humberto
> --
>
> -----------------------------------------------
> Humberto Cervantes <Humberto.Cervantes@xxxxxxx>
> Web: http://www-adele.imag.fr/~cervante/
> LSR-Adele,
> 220 rue de la Chimie, Domaine Universitaire,
> B.P. 53, 38041 Grenoble Cedex 9, France
> Tel: (33)4 76 63 55 74
> -----------------------------------------------
A+
Philippe Merle
--
__________________________________________________
Dr. Philippe Merle - INRIA Scientist Researcher at
Laboratoire d'Informatique Fondamentale de Lille
UPRESA 8022 CNRS - U.F.R. I.E.E.A. - Batiment M3
Universite des Sciences et Technologies de Lille
59655 Villeneuve d'Ascq CEDEX France
Phone: (33) (0)3 20 33 70 24
Fax: (33) (0)3 28 77 85 37
E-Mail: Philippe.Merle@xxxxxxx
Home Page: http://www.lifl.fr/~merle
See also: http://www.objectweb.org/
http://corbaweb.lifl.fr/
__________________________________________________
| <-- Date Index --> | <-- Thread Index --> |
Powered by MHonArc.
Copyright © 1999-2005, ObjectWeb Consortium | contact | webmaster.