Re: [service-orientated-architecture] Re: WebServices to repeat CORBA's mistakes : "The Rise and Fall of CORBA"

Thu, 22 Jun 2006 00:55:39 -0700

Patrick asked:


Do those high-performance scenarios tend to use the same vendor at
all-or-most of the end points, ORBs, etc. Or are there actually a good
number of high-performance scenarios using several vendors
significantly in the critical aspects?

Almost all of them involved using the same vendor at all-or-most endpoints. But that's the way it was with CORBA. It wasn't initially designed to enable interoperability. That was an after-thought, and wasn't a solid reality until after the dual death-knell rang with the introduction of J2EE and the Internet. CORBA died out remarkably quickly -- not because the technology was fundamentally flawed, but because our requirements fundamentally changed. In 1991, most "mission-critical" business application was being developed in C or C++. The Web didn;t exist. IP wasn't even a universal standard. DECnet, SNA, and IPX were still serious NOS contenders. The goal of CORBA was to enable cross-platform integration that could bridge multiple networks. It was a different world.

Interoperable middleware is a product of this century. It wasn't even a dream in 1991. Consider: CICS/COBOL was based on a single vendor, and it's had a pretty successful run. CORBA isn't as pervasive as CICS/COBOL, but it might very well be with us just as long.

Anne


On 6/21/06, patrickdlogan <[EMAIL PROTECTED]> wrote:

> And while the criticisms in the article about the standards process
> are valid in general, and no doubt contributed to some of the
> technical shortcomings in CORBA, standards are not as much about
> technology as they are about agreement and adoption.

The comment about people setting standards they don't know how to
implement seems relevant. Moreover many standards appear to be set
without having been completely implemented by anyone. That seems
antithetical to the term "standard" and aimed more at markitecture
than solving real problems.


> At the end of the day these "my technology is better than your
> technology" arguments often fail to address a key point, which is
> how standards gain adoption.

How standards gain adoption in the IT industry does not appear to be a
process worth being proud of. Sausage making can seem more palatable.


> Well, CORBA did not fail. In fact CORBA is more widely deployed in
> high performance distributed applications than any other
> standards-based technology on the market.

Do those high-performance scenarios tend to use the same vendor at
all-or-most of the end points, ORBs, etc. Or are there actually a good
number of high-performance scenarios using several vendors
significantly in the critical aspects?


> Also whatever criticisms one might have about the Web services
> standardization process, it is a mistake to say that it has an OMG
> behind it. Web services standards are not led by a single
> organization, which is actually a big part of the problem. We have
> OASIS, W3C, WS-I, JCP, and Microsoft/IBM all working on various
> aspects of Web services, also SCA now in fact... This is a very
> different situation than CORBA and the OMG.

No comment needed. I think most of us can reflect proportional
confusion as to the result and future directions.

-Patrick


__._,_.___


SPONSORED LINKS
Computer software Computer aided design software Computer job
Soa Service-oriented architecture

YAHOO! GROUPS LINKS



__,_._,___

Reply via email to