kicad-developers team mailing list archive

[Dick Hollenbeck <dick@xxxxxxxxxxx>]

On 04/20/2014 05:29 AM, Lorenzo Marcantonio wrote:
> On Sun, Apr 20, 2014 at 04:24:18AM -0500, Dick Hollenbeck wrote:
>> Yes globals are global.  Most of them are now a problem, now that I have my mind set on
>> supporting multiple instances of the same KIWAY_PLAYER types in the same process.  I have
>> spoken to Wayne about this and it is on its way into the roadmap as a soon TODO item.
> 
> OK, I guessed right then. Thanks for the explanation.
> 
>> This kind of sorting is really not much more than that, sorting followed by search and
>> replace.  It should not be made into a mountain of work.
> 
> Never said it would. I already removed globals in the past, nothing
> particularly difficult.
> 
> As an aside, why dynamically load the modules and simply don't link them
> in?


The modular-kicad blueprint had been brewing for 4 years within the project and in my
head.  Experience with the project, and in software design in general, told me that this
was the path of least resistance to greatest gain.

Without software design discipline, the ability to craft limited, purposeful interfaces,
you end up with spaghetti.  As an Italian you may like that, but I didn't think it was
right for this project.

In fact, mixing eeschema and pcbnew together in my estimation would have resulted in the
kind of molecular spaghetti that comes about from a two part expoxy mix.  Things would
have initially looked easier, but then gotten a lot *harder* later.

This chosen architecture requires limited, purposeful interfaces to be designed.  That in
itself is worth the educational value to everyone involved.

If you still don't believe, then try and link them together yourself.  That process is in
itself will be more difficult than what you might think, but it could be educational.

Also remember that my vision has python at the top.