W3C home > Mailing lists > Public > www-dom@w3.org > April to June 2002

Re: Level 3 Core Nitpicks: Alternative IDL descriptions and references

From: David Brownell <david-b@pacbell.net>
Date: Wed, 24 Apr 2002 08:50:57 -0700
To: Steve Barber <sbarber@randomwalk.com>
Cc: www-dom@w3.org
Message-id: <12f201c1eba7$d3dfa640$6800000a@brownell.org>
... except that while MSFT's IDL (used with COM) was designed
for a particular computing environment (C/C++ oriented, look
at it in detail), that was explicitly a non-goal of OMG-IDL.  In fact
one of its design goals (one that was achieved!) was getting past
some of those kinds of problems with DCE-IDL (from which
MSFT borrowed heavily, ISTR :).

It'd be worth saying that of the object-oriented IDLs in wide use
at the time that DOM Level 1 was specified , OMG-IDL was the
least bound to a specific computing environment.

- Dave


----- Original Message ----- 
From: "Steve Barber" <sbarber@randomwalk.com>
To: <www-dom@w3.org>
Sent: Tuesday, April 23, 2002 11:42 AM
Subject: Level 3 Core Nitpicks: Alternative IDL descriptions and references


> Apologies for the nitpicks, but since the DOM Level 3 will likely become a
> Recommendation, may as well be precise.
> 
> In http://www.w3.org/TR/DOM-Level-3-Core/introduction.html is the following
> Note:
> 
> "Note: OMG IDL is used only as a language-independent and implementation-
> neutral way to specify interfaces. Various other IDLs could have been used
> ([COM], [Java IDL], [MIDL], ...). In general, IDLs are designed for
> specific computing environments. The Document Object Model can be
> implemented in any computing environment, and does not require the object
> binding runtimes generally associated with such IDLs."
> 
> The same language is in the DOM Level 2 Core (http://www.w3.org/TR/DOM-
> Level-2-Core/introduction.html).
> 
> Proposal #1:
> 
> Replace the second sentence of the Note ("Various other IDLs could have
> been used ([COM], [Java IDL], [MIDL], ...).") with:
> 
> "Other IDLs such as [MIDL] could have been used."
> 
> Delete the Java IDL entry from I.2 ("Informative References").
> 
> Delete the COM entry from I.2 ("Informative References").
> 
> Rationale #1:
> 
> Java IDL is not a different IDL than OMG IDL. JavaIDL is Sun's
> *implementation* of a CORBA-compliant ORB that is compatible with some
> level of OMG IDL.
> 
> COM is arguably not itself an IDL. MIDL is the IDL used in COM, and MIDL is
> mentioned elsewhere in the note.
> 
> Proposal #2:
> In I.2 ("Informative References"), the URL given for MIDL no longer works.
> I propose replacing it with one that does. Here's one that worked for me,
> today: http://msdn.microsoft.com/library/default.asp?url=/library/en-
> us/midl/mi-laref_1r1h.asp
> 
> Comment:
> If there's a need to list more than one alternative IDL, there are others.
> ONC RPC and DCE RPC come to mind, though these are not completely object-
> oriented. For a generalization of the IDL concept, see
> http://www.cs.utah.edu/flux/flick/
> 
> -- 
> Steve Barber
> Senior Consultant
> Random Walk Computing, Inc.
> sbarber@randomwalk.com
> Co-author, Programming with JavaIDL (Wiley 1997)
> (http://www.echonyc.com/~sbarber/javaidlbook/)
> 
> 
Received on Wednesday, 24 April 2002 12:10:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 June 2012 06:13:56 GMT