Mail Archives: geda-user/2015/12/30/00:44:44
On Tue, 29 Dec 2015, John Doty wrote:
>
> On Dec 29, 2015, at 9:54 PM, gedau AT igor2 DOT repo DOT hu wrote:
>
>> - gschem doesn't have cosistent concept of its goals. It pretends it doesn't need to know about nets because it's a dump editor, but it does know about slotting and has lists of hardwired attribute names in code
>
> And some of us think that those are design errors. There are several alternate notions of how slotting should work. If you want slotting in gschem, you should load a suitable script.
You are trying to avoid answering the real problematic part, picking on
the small things.
Any EE I've ever met or worked with talked the same language. The most
important bricks of the language were compoment and network names. Does
gschem speak this language?
Compontent names: refdes; it's hardwired all around to make gschem more or
less understand the language. Still it is not strong enough so that we can
really idetify (search, find, list) a component by name and make sure we
get what we need to get.
Network names: gschem has absolutely no idea about networks. Gschem fails
to model this aspect of the world.
I am all for plugins and scripts (but against restricting the user to one
specific scripting language). However, there must be a core of the tool,
which is not a plugin, not an user provided addon. This core should try to
speak the same language to at least a minimal extent, that the userbase
does. To me it seems gschem fails on this. It probably could be worked
around with endless streams of scheme scripts, but it'd be easier to admit
it and come up with a proper fix in the core.
- Raw text -