🚀 go-pugleaf

Dean Schulze wrote:

>     Bob, I don't mean to diminish the urgency that you bring to this
> issue,
> but didn't OpenDoc die because the two companies who were its main
> advocates
> (Apple and IBM) were both laggards when it comes to introducing new
> technology
> to market, and especially when it comes to understanding the importance
> of
> getting independant developers to support their standard?  Also, didn't
> JavaBeans have a lot to do with the demise of OpenDoc?  How about
> Apple's
> financial problems?

Dean, you make good points about why OpenDoc died.  What I was trying to say is
that the ORB is a key piece of the battle for the 3-tier Object Web.  The other
key pieces are the component model, component coordinators, tools, and so on.  In
the case of the Microsoft Object Web the answer is simple: ActiveXs on the client,
Viperized ActiveXs on the server, and DCOM for the distributed object glue.  This
is all supported by an incredibly powerful set of tools (and product suites).

For our side to provide the same thing, we must have JavaBeans, Enterprise
JavaBeans, JavaBean containers, JavaBean transaction coordinators, great tools,
and an ORB that pulls it all together.  It will take a miracle for all these
pieces to line up.  It will require very key technology alliances between JavaSoft
and the rest of the distributed object community including--the ORB vendors,
Netscape, BEA, IBM/Lotus, Oracle, Informix, Novell, Sybase, and the tool vendors
(Symantec, Borland, etc).  Some of these alliances are coming together well.
However, some of it is of to a very bad start--for example, the RMI vs CORBA
stuff, JNDI vs CORBA naming, etc.  JavaSoft is creating an impedance mismatch with
the rest of the Object Web community.  The ORB is a key piece because it is used
to integrate the pieces.

So why are they doing this?  It could be because of internal Sun politics (very
likely).  It could be because they are not aware of the bigger issues (unlikely).
It could be that they think they can single-handedly beat Microsoft (perhaps).  In
any case, I don't think it's a good idea.  If I wanted a one-shop solution, I
would get it from Microsoft (it's the safer bet).  On the other hand, if the open
coalition ever comes together it can be unbeatable.  But it must no be a
coalition of press releases--it must be based on technology.  The good news is
that all the key players of the open coalition--except for JavaSoft--are 100%
behind CORBA (and this is starting to show in products).  JavaSoft is 50% CORBA
and 50% RMI.  :-)

So what does this have to do with OpenDoc?  OpenDoc was our last attempt to field
an open component architecture against ActiveX.  However, the pieces never came
together.  There were no tools.  There were no TP Monitors for components.  And
there were no portable parts.   And they adopted Java too little too late.  These
are some of the things we warned about in our book "The Distributed Object
Survival Guide"--over 18 months ago.  At that time we wrote a book, pointed out
the problems,  and then hoped someone would fix them.  Well, they didn't.  So this
time we decided not to leave anything to fate.  We wrote a book, pointed the
problems, and we are now making sure they get fixed.  I've seen too many of my
friends (small developers) get hurt because the technology stars couldn't line up.
I don't want to see this happen again.  We must be proactive this time.  JavaBeans
with CORBA are our last chance at fielding an open Java-based component
architecture that can stand up against ActiveX.  Let's make sure our vendors
provide us with all the pieces.  If not, then let's scream.  If they still don't,
then it may be time to make our peace with Microsoft.  Incidentally, Microsoft is
showing a lot of vision in the area of distributed components.  They also
understand the Object Web and distributed objects.  And, they only have one ORB
(and it's free).  In contrast, Sun originally had four ORBs planned for the JDK:
RMI, Doors, NEO, and Joe (all internal Sun technologies).  Now they're down to two
:-)  So there's still hope.

Bob Orfali