[asterisk-dev] Improvement of coding practices

[SF Markus Elfring]

> What do you mean?
>   
I imagine that other code parts will also benefit if the issue "reload 
does not initialize configuration values" will be fixed.

> The problem is that you don't really know that you changed the
> behaviour. At least with a complex code.
>   
Are those APIs clean enough to play with different implementations 
behind the programming interfaces to check for undesired effects?
What is the current state for the separation of concerns?

> No. We're talking about devices that spit at least 1000 interrupts per
> second. Not something that works well using simulations.
>   
I would not say that such an engineering approach is impossible.
Which software and hardware would be able to handle this kind of 
simulator? (I assume that the requirements may be not so overwhelmingly 
high to try it on an ordinary personal computer or a more powerful 
workstation.)


> Why? A patch to one does not touch another. So no need to split just to
> allow concurrent work.
>   
Source code control systems like subversion hide this to some degree.
But this reaction does not fit to complaints about file sizes that are 
too big from a design viewpoint.


> The code today is incomprehensable.
We all know that it is a lot of work to turn it into high-quality 
software pieces that are comprehensible.

Regards,
Markus