On Fri, 2005-05-27 at 00:41 -0400, Daniel Carrera wrote:

> In the sense of borrowing code, the answer is no. Not possible. Not even 
> a little bit. It is very rare when you can grab two large programs and 
> copy code between them and have the thing work. And OOo and Mozilla use 
> completely different component models and all. It would likely be easier 
> to just make a new program.

I think the whole idea of producing megalithic software edifices is a
bad idea in principle. Here are the reasons.

As long as there are agreed interoperability standards and common user
interfaces to the user what is and isn't a separate program is
immaterial. If you ask most lay users they don't know the difference
between the OS and an application. In fact the OS is a collection of
interacting programs with varying degrees of dependency.

The more self-contained applications are the less likely they are to
bring down the entire system if something goes wrong in one component. 

Its easier to maintain and develop discrete apps rather than one
gigantic one where making a change in one place can have unforeseen
consequences somewhere else. 

Its easier to promote competition if improving something is simply a
matter of contributing to one app which has simple and published
protocol for interacting with the rest of the system.

The main reasons for large scale mega projects are

Having all the common parts in libraries makes life easier as each
program does not have to replicate commonly used routines. This does
produce dependencies on certain library modules being in place and even
certain versions. These libraries can become large and might even
contain routines you never use in a particular application.

If you want to port software between platforms common code that can
easily be ported helps but tends to be another constraint on the modular
approach.

If you don't envisage other people contributing to your project it might
well be a commercial advantage to control a large amount of completely
interdependent code because it effectively locks out anyone else.

Since many large FLOSS projects such as OOo started life as closed
source there is a legacy which I doubt we would consider desirable if we
started from scratch. The whole point of multi-tasking operating systems
is that you can run several independent applications concurrently and
cooperatively and pass information between them as if they were one
application. That way an Office suite could actually mean all the things
used in an office, it would just mean that you could mix and match
components and not necessarily get all the software from the same
source. Defining an open document format based on XML is a step forward
in this respect because it encourages third parties to develop
applications to integrate with each other. So I believe we should resist
just adding more and more features to OOo especially when there are
perfectly good tools out there freely available to do the job.

> Trying to make them look similar is a fair goal. 

And in the whole scheme of things both possible and desirable. Distros
already do this to an extent in the Linux world because they have the
flexibility to and I'd expect that to keep developing. 

> But cosmetic similarity 
> is nothing at all like borrowing code. They would still be distinct 
> programs.
> 
> Cheers,
> Daniel.
-- 
Ian Lynch <[EMAIL PROTECTED]>
ZMSL

Reply via email to