httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Behlendorf <>
Subject Re: wild prediction
Date Mon, 03 Jun 1996 04:43:29 GMT
On Mon, 3 Jun 1996 wrote:
> > I'd throw in OpenDoc, assuming Apple and co. can really get its act
> > together soon, as it claims it can. One advantage of OpenDoc is that
> > it can already use OLE peices, and within the year (so they claim)
> > it'll have Java built in too. Which takes care of 2/3 of your list.
> > And while I don't have the hardware (here at home) to try it, people
> > keep claiming Cyberdog (Apple's OpenDoc-based Internet architecture)
> > is "really cool".

Remember, technology alone is not going to designate the winner.  OpenDoc 
might be superior, it's certainly a bit more mature, even though the 
OD-based browser is just coming out.  However, I just don't see Apple 
having the ability to pull contenders in the industry together to "agree" 
on it - interoperability is one thing, but Sun's convincing Microsoft 
that building a really fast JIT compiler for MSIE 3.0 is something I 
don't think Apple has the comparable ability to do.  

> Then I'll counter with Corba.  

At the "Distributed Code and Persistance" session at javaOne, the head of the
"javaSpace" and java-to-java remote method invocation development efforts Jim
Waldo was asked his opinion of Corba.  He said something to the effect of "I
don't want to turn this into a Corba-bashing session, honest - I was involved
with Corba from the start, I set up I believe the first ORB, and I can say
I've watched corba evolve from a simple protocol into a very complicated
protocol that takes a lot of implementation effort, and it's sorta like
seeing your kid out on the street in fishnet and lingerie and thinking, damn,
he could have been somebody!" 

> There is already a free implementation (ILU)
> along with a Java ORB (Jylu).  And Orbix's commercial products are
> truly impressive.  Corba, OLE and OpenDoc should actually be able to
> get along and I am not convinced that we are going to have a clear winner
> here.  Everybody can keep doing their own thing and with a little glue
> the various brands of clients will all talk to the various brands of
> servers.

The thing is stable markets are not necessarily made of competing 
equivalent platforms.  Imagine if there was another protocol that did 
pretty much exactly what HTTP did, with the same amount of (in)stability 
and performance, with a free implementation - would HTTP have survived or 
would this other-protocol?  When the technology is very close, psychology 
comes into play, and whichever technology appears to have the slightest 
lead (like, one side has a group solidly supporting a public-domain 
server and extending it, etc) then the weight of the world would fall to 
the other.  However, when the technologies are distinctly different, they 
can coexist... if OpenDoc were to survive it'd have to distinguish itself 
from other technologies pretty strongly.  It's not a VM, it's not a 
remote procedure call technology, it's a compound document and linking 
technology (right? or are they biting off more these days?)


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--  |  We're hiring!

View raw message