OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   RE: Alternatives to browsers (was Re: Alternatives to the W3C)

[ Lists Home | Date Index | Thread Index ]
  • From: "Didier PH Martin" <martind@netfolder.com>
  • To: "Tim Bray" <tbray@textuality.com>,<xml-dev@ic.ac.uk>
  • Date: Tue, 18 Jan 2000 00:19:09 -0500

Hi Tim,

Tim said:
Nobody would call it preferable in the general case.  The story I keep
hearing is that in comparison with XML, HTTP, string, and glue, frameworks
like CORBA and RMI offer you, say N times as much (security, transaction
semantics, etc), but cost M times as much time & money to deploy.  In some
application contexts the relative values of M and N would predispose you
toward CORBA, in others towards using XML.

If there's an underlying lesson, it's that the Web is all about doing
a lot with a little; the HTML/HTTP/URI trio have to count as one of the
great 80/20 point bullseyes in the history of technology.  With XML
messaging and HTTP piping, you can sometimes do a whole lot remarkably
quickly.

Doesn't mean CORBA's obsolete.  DCOM, maybe. -T.

Didier reply:
Does not mean that DCOM is dead either. In fact, the good question to ask is
what is CORBA and DCOM after all, and is the difference between CORBA and
DCOM?

Q: what is CORBA and DCOM after all?
A: CORBA: an interface definition format named IDL. a marshalling convention
named IIOP. The notion of encoding-decoding with the usage of stubs (that
encode and decode from programming language into the marshalling format).
DCOM: an interface definition format named IDL but different from CORBA (in
fact based on a previous consortium name OSF). A marshalling format based on
the OSF RPC marshalling format. the notion of encoding-decoding with the
usage of stubs (in fact named in this case proxy and stub - both used to
encode and decode from the marshalling format into a programming language).
CORBA is available as free software (IDL compiler, run time ORB etc.) and as
product. It is, I also bundled with the Gnome package and any Linux package
that include GNOME.
DCOM is bundled in the 150 millions copies of Windows and is also available
(bug yet buggy) as free software (the WINE project). Could also be obtained
as free software for the Unix/Linux platform (re: Software AG) and available
as product.

Q:What is the difference?
R: the IDL and how objects/interfaces are structured.
DCOM includes a binary interface definition and thus allows true
inter-operability between components even created with different languages
to work together. Thus, components could be used intra-address  space
(without marshalling) or inter-address space (with marshalling)
CORBA only include an inter-address space mechanism and has no binary
interface convention.

Q: what would be the best world?
That both CORBA and DCOM provide an IDL to an XML based marshaler. This way,
at least, two components (a) a CORBA one and (b) a DCOM one could be made to
inter-operate if both also use HTTP as a transport mechanism.

Will DCOM disappear? Come one Tim, get out of your home :-) How to make
disappear 150 million windows just to say dip a dip doo? and bang it all
disappear with a touch of a magic wand? So DCOM is here to stay as long as
window as a comfortable market share.
Will CORBA increase its market share? yes if all Linux package bundle the
same ORB. Will CORBA get a bigger market share than DCOM? only if all Linux
packages does (a) bundle the same ORB, (b) increase their market share. (b)
is happening, (a) is not, the Linux world is not an homogeneous one, there
are several resellers and packagers that obviously are trying to
differentiate and play, in certain ways, the same game as Microsoft does
(try to get their customers captive).

What is the advantages of both DCOM and CORBA?
have an interface definition language that allows you to compile it and
create a mapping for a particular language. You then created a meta language
for procedural or non procedural languages. It resolves the impedance
mismatch between components and allows the development by components instead
of re-inventing the wheel each time.

What W3C could do to really help us create inter-operable apps?
forget about the IDL and crusade about CORBA. Just focus on the marshalling
format and do a recommendation for the marshalling format and transport
mechanism. Get in the WG representative from Microsoft, the OMG consortium
and the Linux community (are there any Linux packager part of the W3
consortium?) and get them agree on a common marshalling and transport
format. Believe me, we'll have accomplished a lot toward inter-operability
of applications.

What is in the way to make that happen?
Simply the will of somebody to put that on the W3C agenda.

Cheers
Didier PH Martin
----------------------------------------------
Email: martind@netfolder.com
Conferences: Web New York (http://www.mfweb.com)
Book to come soon: XML Pro published by Wrox Press
Products: http://www.netfolder.com



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ or CD-ROM/ISBN 981-02-3594-1
Please note: New list subscriptions now closed in preparation for transfer to OASIS.






 

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS