Re: Which way to the Object Web?

Ron Resnick (
Mon, 14 Sep 1998 11:09:09 -0400

Jeff Bone wrote:
> Query for everybody. I've been taking a re-look at CORBA after not
> paying much attention to it for the last couple of years... seems to me
> that the specs have really evolved, that the product maturity has
> improved dramatically, etc. Given all that, an old quandry has been
> knocking around in my head again...
> CORBA &co vs. HTTP + extensions + XML &co?
> Anyone have any technical, political, philosophical, or other thoughts
> on this? Love to hear 'em...
> jb

CORBA big blech. What, you no hear Java? Jungle you live in which?
Rock behind hiding where?

Seriously, distributed objects <> CORBA.
If objects+web is what you crave, there's much more going on than
CORBA. Most of the fun stuff in Internet-based objects is happening
in distributed Java, imho. Sure, there's Java CORBA orbs
(Iona, Inprise),
but the neater stuff - mobility, servlets, loosely coupled
message buses etc. ain't there, it's
in Voyager, various webservers, Tengah, Javaspaces, JSDT,
iBus and other such animals.
I think much of the clueful distributed-objects
research, experiment, and product has moved off of CORBA and
increasingly onto Java.

True, there seems to have been a resurgence of commercial interest
in CORBA in the last 12 months or so. Even here though,
CORBA-circa 1998 has a new name: EJB (enterprise Java Beans,
which has a lot more to do with CORBA than with Beans, afaict).

You think the corba specs are displaying maturity? Which ones?
You think the products are matured? Not the ones I use
(primarily visbroker, some orbix). I'm a lot
more impressed with some of the freeware stuff (TAO, omniORB2, Electra)
than with the most of the "product". OK, that's just a general
statement in favour of open source, but it really seems to apply in
CORBA space.

Watch for Java/XML as one of the key emerging trends in the next
12 months for "the middleware that glues it all together". Sure,
we've all heard that refrain a dozen or more times (EDI, COM, DCE,
OSI, CORBA, Web, blippity-blippity), and no doubt this one will fall
short too (it pretty much has to - that's preordained). But on
each bounce off the trampoline, we learn a bit more, and leap
a bit higher. Java/XML efforts are the next big bounce. CORBA
is a past-tense, been-there done-that, bounce.

Want a bit more detail?
Try the "Large Scale Distributed Java" ppt presentation I have up at

[And once I'm handing out that url, let me also plug
"Distributed Object Management in the Large:
A Fluid Model for Object Ensembles" at the same site - html format.
Particularly interested in comments from folks with a background
in both physics & conservation equations, as well as distributed

If you want informed discussions of where largescale distributed object
computing is going, regardless of whether it's corba, com, java, web
or what, you may be interested in

Or, if you *really* want to learn the inside skinny on corba,
where it's been and where it's going, my company (DiaLogos)
offers Java & CORBA training -
That's my day job - teaching IDLs and BOAs and structures of
sequences of interface types & language mappings to folks who
think it's all the latest advanced stuff. Ya gotta start somewhere!

Naturally, my corba sentiments here are personal comments only,
and do not reflect anyone else, particularly not DiaLogos.