Re: Is CORBA dead for Ada

[Victor Giddings <victor.giddings@ois.com>]

Laurent Pautet <pautet@enst.fr> wrote in news:bn145s$1kbk$1@avanie.enst.fr:

> It seems to me that the CORBA mapping for Ada is almost dead. The last
> official document is formal/01-10-42 for CORBA 2.3.
> 
> The ada-rtf team does not seem to be very active in its job to update
> the mapping (the last closed issues are from 1999).
> 
> OIS which seems to lead this task force is very shy in its promotion
> of Ada on its web site.

I will feed this back to our marketing team. We are not shy in our 
promotion of Ada. In the last OMG Real-Time Workshop in response to a 
comment, I was accused of being "an unrepentent Ada programmer". At which 
point, my boss exclaimed "damned right we are"!

> TopGraph'X is still promoting its Ada products. If ORB-River is
> compatible with CORBA 2.6, it seems that it does not include new
> features from versions greater than 2.3.
> 
> Does anyone have fresh news on the OMG activities around the Ada
> mapping ?
> 
> --
> --  Laurent

Dirk Craeynest was kind enough to repost my responses to a similar thread 
on comp.object.corba and inform me of this thread. I would like to expand 
on some of the earlier responses. 

There seems to be much apprehension and more than a little misunderstanding 
associated with CORBA "versions". CORBA is not a monolithic specification 
but a collection of adopted specifications that may or may not be 
consolidated into individual documents. The individual documents have 
associated Revision Task Forces (RTFs) and individual life cycles. The 
references such as CORBA 2.6 are more properly references to a particular 
version of the "CORBA Core" specification, a particular document that 
specifies the language-independent requirements of what an ORB product must 
implement. Changes to the CORBA Core specification may or may not require 
changes to the language mapping specifications, depending on whether there 
is a significant change to the IDL language. Therefore, a lot of revision 
of the CORBA Core would have been addressed by changing the last digit in 
the statement that the mapping was "aligned to CORBA version 2.x". 

There is understandable confusion about this that has been partly caused by 
the OMG itself. First of all, the CORBA Core document contains a lot of 
things, e.g. CORBA/COM Interworking, that don't have to be implemented by 
an ORB product. The OMG staff have also issued press releases that claimed 
what the contents of CORBA x.x would contain. In general, this is a problem 
that needs to be fixed. In addition to being the chair of the Ada RTF, I am 
chair of a group in the OMG called the Product Specification Definition 
(psdef) subcommittee that is trying to straighten out publication 
organization and coordination of versioning. I urge you to participate. 

The bottom line is that the fact that the current Ada Language Mapping 
specification is "aligned to CORBA 2.3" means very little. As I stated in 
the comp.object.corba post, the OMG IDL language is fairly stable, so few 
changes in the language mapping are needed. Most of the features added in 
the CORBA Core 2.4 and later versions are specified in a language 
independent manner and have not affected the language mapping. So, as Jean-
Claude Mahieux was able to report, there has been no hindrance to advancing 
Ada ORB implementations. 

That being said, the other responders are correct in that there has not 
been an active Ada Mapping RTF is almost 3 years. My only excuse for this 
is that we have been busy with other OMG specifications like Real-Time 
CORBA (1.0 and 2.0), Fault Tolerant CORBA, Data Distribution, etc., etc.. 
(I'm not sure what excuse my competitors have ;) Nevertheless, there is 
currently an active RTF attempting to deal with the backlog. Anyone may 
participate in this activity by joining the email group (ada-rtf-
request@omg.org) Non-OMG members may need to contact me to be added to the 
list. Voting membership (one per organization) in the RTF requires a 
minimal level of membership in the OMG but, in practice, most voting is 
pro-forma after a consensus has been worked out among us. Let me know if 
you are interested in formal membership, so this can be placed on the PTC 
agenda. 
 
-- 
Victor Giddings		mailto:victor.giddings@ois.com
Senior Scientist	+1 703 295 6500
Objective Interface Systems	Fax: +1 703 295 6501