Re: Thoughts on Design Principles was Re: [Gnewsense-dev] Report of non-free software in gNewSense

[Karl Goetz]

On Tue, 09 Jun 2009 19:18:11 +0100
Paul O'Malley - gnu's not unix - <address@hidden> wrote:

> > 

> 
> the success of gns has been driven by making it easy to repeat the
> steps forward - maybe it is time to expand this

I think it is time. several of us have started talking about ways to
improve the processes used, but for the next few weeks I don't really
want to start this discussion properly.

> When we initially sat down to design the way this system was built we 
> did the following.
> 
> Rule 1:  anything you do must be easy to repeat

I'd like to keep this /to some extent/, but I'm not sure the way we do
it currently is how I would like things set up.

> 
> Rule 2:  remember the changes you make increase the distance to your 
> source and the work you must do for little return

This is true. However, if we change our processes (to hopefully make it
easier for people to work with us), perhaps we can increase our deltas.
Changing how we make our changes might even allow us to contribute back
to upstream projects in more then just a 'we removed this' way.

> 
> The smaller the delta to upstream the easier it is to maintain.
> If you change one thing make sure you are willing to keep maintaining 
> that change.

Happily for us, the issue of re-merging all our modification to upstream
packages once our code is working only happens every few years.

> 
> This informed our lack of changes and importing, had we more support
> we would have been open to changing modality.
> 

Hackers seem to join the project slowly, but that might be in part due
to our system (I dont know, its just a thought I've been carrying. I'm
also not fooling myself that changing our systems will magically
produce contributions).

> People are working hard on Lemote and I am wholly trilled to see this.

+++++

> 
> Perhaps it is time to look at a design of a system build differently.
> 
> Joining other distributions seems attractive at until you get up
> close, build methods are not well defined in my personal opinion, or
> they are not systems that are suitable for general use.
> 
> So where is this all going. You tell me! ;-)

Is it alright for me to ask this conversation to restart in a few
weeks? :)
kk

> 
> Regards,
> Paul
> 

-- 
Karl Goetz, (Kamping_Kaiser / VK5FOSS)
Debian contributor / gNewSense Maintainer
http://www.kgoetz.id.au
No, I won't join your social networking group

signature.asc
Description: PGP signature