----- Original Message ----

> From: Charles Marcus <[email protected]>
> On 2010-11-30 5:29 PM, BRM wrote:
> >> LibO - like OOo - does not really  have separate components. Even  if you
> >> > could download just  one component, the resulting size would only be  a
> >> > few MBs  smaller than it is now.
> > While that may currently be the case - that is  absolutely ridiculous.
> > TDF/LO should make a priority of resolving that  issue.
> 
> Great, then I'm sure your contribution of these code changes will  be
> forthcoming soon?
> 
> Yes, I'm joking.
> 

As joking as you may be, I for one would do so if I had the time.
As it is - I might in a few months, but I can't guarantee it right now.

The cost/benefit would _be_ worth it.

> As was explained to me -  it is not just that it is a huge job, it is a
> monstrous job that would  essentially require rewriting the entire code
> base from scratch. I have also  heard horror stories from very
> experienced programmers when they start  getting into the code.
> 

And that is exactly why it would be worth it. If experienced developers are 
having trouble
getting into the code, then just imagine how much the code itself is turning 
people away
from contributing to LO/OOo. Resolve that and you will likely get a lot more 
contributors
of varying experience levels - or at least, more experienced contributors.

> So, again, no one is saying this wouldn't be a  wonderful thing, it is
> simply not something that is feasible at the  moment.

It will never be feasible. So instead of saying "yeah that needs to be done, 
but 
its not feasible" let's hash out a plan for actually doing so.
 
> > Having created installers before - namely MSI's - there  should 
> > ultimately be no reason why the installer should be broken down  as:
> > 
> > - core LO libraries used by each package
> 
> And again,  the point is, these core LO libraries used by each (assuming
> you mean  used/shared by all packages) consist of 98% of the size of the
> download,

I mean:

- static and shared libraries that consist of code utilized by independent 
executables for each program.
- ancillary programs that aid each program
- etc.

>  so  the cost/benefit ratio is just not worth it.

And again - you made my exact point that the cost/benefit will be worth it if 
not for getting more contributors alone.

Please stop discouraging this kind of work. If the effort is to be done at all, 
then we need to encourage this kind
of work - even if in small incremental steps. But it has to start somewhere and 
with a goal in mind to accomplish.

Are the LO/OOo developers/management that averse to change? I certainly hope 
not.
Is TDF that adverse to change? I certainly hope not as well.

Ben



-- 
Unsubscribe instructions: E-mail to [email protected]
Archive: http://www.documentfoundation.org/lists/discuss/
*** All posts to this list are publicly archived for eternity ***

Reply via email to