1. OOP vs Messaging (was Indigo on Wintel)
The problem with OOP in inter-program communication is that you can't
transmit an object as it exists in a program between programs. You can
perform serialization, as it's called in Java, but what you get is a coded
byte string that can be used to recreate an object. This has always been
problematic, since the reconstruction requires the receiver of the byte
string to have access to the same class files as the sender. Not unmanageble
when the communication is among multiple instances of the same program, but
a nightmare in the more usual cases.
Microsoft's decision to use messages (in this case XML documents) is a sound
one. It should avoid many problems that occur in DCOM, RMI, CORBA, etc.
This does not mean OOP is useless, nor does the article assert that it is.
OOP is a useful technique in the construction of programs. Nor is
serialization useless--it shines at its real purpose, which is allowing an
OOP program to easily save and restore its own state.
Concluding that OOP is not a good tool because it works poorly in
inter-program communication is like concluding that a hammer is no good for
driving nails because it is useless for sawing boards.
-- Mike Nelson
2. Re: OOP vs Messaging (was Indigo on Wintel)
Mike Nelson wrote (some snippage occurred):
>The problem with OOP in inter-program communication is that you can't
>transmit an object as it exists in a program between programs...
>
>Microsoft's decision to use messages (in this case XML documents) is a sound
>one. It should avoid many problems that occur in DCOM, RMI, CORBA, etc.
>
>
Mike, I'm glad you chimed in because of your expertise with OOP. All
that is happening is that OOP programs will no longer "communicate with
objects" (if they ever did) but with messages... right? And I agree that
seems to be a good thing. :)
>Concluding that OOP is not a good tool because it works poorly in
>inter-program communication is like concluding that a hammer is no good for
>driving nails because it is useless for sawing boards.
>
>
I wish I had time to devote to understand OOP enough to understand if I
want to go that route... ;)
Thanks!
-ck
3. Re: OOP vs Messaging (was Indigo on Wintel)
----- Original Message -----
From: "Mike Nelson" <MichaelANelson at WORLDNET.ATT.NET>
To: <EUforum at topica.com>
Subject: OOP vs Messaging (was Indigo on Wintel)
<snip>
> Microsoft's decision to use messages (in this case XML documents) is a sound
> one. It should avoid many problems that occur in DCOM, RMI, CORBA, etc.
>
<snip>
>
> Concluding that OOP is not a good tool because it works poorly in
> inter-program communication is like concluding that a hammer is no good for
> driving nails because it is useless for sawing boards.
I'm in total agreement, Mike.
--
Derek
4. Re: OOP vs Messaging (was Indigo on Wintel)
On 29 Jan 2004, at 7:30, Derek Parnell wrote:
>
>
> ----- Original Message -----
> From: "Mike Nelson" <MichaelANelson at WORLDNET.ATT.NET>
> > Concluding that OOP is not a good tool because it works poorly in
> > inter-program communication is like concluding that a hammer is no good for
> > driving nails because it is useless for sawing boards.
>
> I'm in total agreement, Mike.
OOP's way of doing things included it's methods of communication. That's
the problem i have with it. Encapulating data and program code can be done
different ways, even in a mirc database. That's why M$ is moving away from
it. Or did imiss something (regardless of if M$ can gain more ownership
control over whatever they move to)? Otherwise, OOP isn't that much
different than over-hyped functions. Same as a vxd isn't any more than a over-
hyped dos device driver (click different buttons to compile differently).
Kat
