Re: Distributed programming, a killer application for ADA ?

["Nick Roberts" <nick.roberts@acm.org>]

"Warren W. Gay VE3WWG" <ve3wwg@cogeco.ca> wrote in message
news:pxHxc.28025$8k4.722102@news20.bellglobal.com...

> >>>Perhaps demonstrating how to write concurrent distributed
> >>>programs with Ada might help to create interst.
> >>
> >>Nobody but ACT supports the distributed annex
> >>Apparently they think that CORBA is better.
> >
> > Oh. :-(
> >
> > Sorry to hear that.
>
> What gives rise to "apparently they think.."? It is
> best not to read too much into impressions. They may
> in fact have customer demand for CORBA, but
> that doesn't necessarily mean that the distributed
> annex has fallen out of favour.

I'm certain that specific customer demand for CORBA is one big reason.

However, be aware that the DSA (Annex E) and CORBA are not direct
alternatives for one another. It is a crucial aspect of CORBA that it
supports programs developed in isolation from one another except for the
interfaces between them (defined, actually or in effect, by OMG IDL); in
particular, multiple programming languages can be used. The DSA requires
that all the partitions of a distributed program are compiled in the same
environment at least to the extent that every partition has a means of
identifying the others it depends on; CORBA has no equivalent requirement.
The DSA requires (but does not specify) a mechanism by which all the
partitions come into execution simultaneously (in effect). CORBA makes
provision (specifying some, but not all of of) for much more flexible
mechanisms, for example supporting the execution of server programs upon
demand. There are many other significant and fundamental differences.

(Incidentally, I intend to support /both/ DSA and CORBA in AdaOS,
eventually.)

-- 
Nick Roberts