[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: Common server activation, the sequel
From:       Bavo De Ridder <bavodr () poboxes ! com>
Date:       1999-06-28 14:53:34
[Download RAW message or body]

On Mon, 28 Jun 1999, Simon Hausmann wrote:
>On Mon, 28 Jun 1999, Bavo De Ridder wrote:

>> 
>> What we proposed was the following:
>> 
>> (1) define a common corba component type (something like KOM::Base...). Every
>> future component should inherit from this base component. This base component
>> differs from KOM::Base... in that it is meant for server-side instantiations.
>> This probably means that a few extentions should be provided. We can off course
>> reuse the signal/slot mechanism through inheritance.
>> (2) build and define a generic server. This application's sole purpose is to
>> offer an execution environment (for the non-corba guys: an address-space to run
>> in). At the same time the container-server (or in more marketing like words: an
>> application server), will also offer a standardized access to generic services
>> like events (signal/slot, COS Event Service), naming services, persistance
>> services, ... Using the POA we can also offer thread-safe subcontainers
>> (apartments in COM).  
>> 
>
>Having a common standard like described in 1) and 2) is IMHO unlikely to
>happen, since it would require use to share object models and other core 
>interfaces.
>

It is not unlikely to happen. The common standard already exists. Have a look
at the CORBA Component Specifications, available on the OMG site.

The thing you have to agree on at least is some common IDL, even for core
services. Unless you want to scare away independent software developers. They
(the independent software developers) just think: how can I write my
editor-component so it will run on both KDE, Gnome, ... . Currently this isn't
possible. They are three options: 

(1) Gnome and KDE agree on a common IDL
(2) They both go their own way, this will end up in a civilar war on the
internet between Gnome and KDE in which only end-users will suffer. In the end
one will win but Linux will come out weakened compared to Microsoft. Bill wins
this one
(3) The same as (2) but nobody wins. This will destroy KDE/Gnome in the long
run and a third (probably commercial) desktop will win, and guess what, I have
a strong feeling that commercial desktop will be something Microsoft will bring
us.

It is up to you, KDE and Gnome developers, to decide ...

Sidenote:

Both KDE and Gnome heavily suffer from re-inventing the wheel:
they both re-invented several CORBA specifications: CORBA Component Model,
naming service and recently I read postings about throwing the C++ binding away
in favor of a Qt based binding or stripping down the ORB.

IMHO this is destroying what made Linux great: standards and compatibility.
This is *not* the way to go !!!!


BDR

-------------------------------------------------------------
Bavo De Ridder <bavodr@poboxes.com>

PGP Public Key: http://www.poboxes.com/bavodr/public.key.pgp
-------------------------------------------------------------

[prev in list] [next in list] [prev in thread] [next in thread]