Mail Archives: geda-user/2015/09/22/18:20:58
On Tue, 22 Sep 2015, Markus Hitter (mah AT jump-ing DOT de) [via
geda-user AT delorie DOT com] wrote:
> People currently have an extremely strong tendency to rewrite everything
> from scratch instead of looking at and improving what's already there.
> As you did yourself.
That's not true. I started out trying to improve gEDA. Only after I
found fundamental problems which I couldn't realistically solve by
improving the existing code, I started writing something new. Also, I
definitively didn't "rewrite everything from scratch"; on the contrary, I
tried to preserve as much existing code as reasonable.
This isn't always obvious as I moved many things around and renamed them,
and sometimes much code collapsed into a few simple functions; but if you
look, for example, at slib.py, clib.py, or attrib.py, you should be able
to tell which code corresponds to which libgeda code.
The only things I wrote from scratch are libxornstorage, its Python
bindings, and the command-line interface; almost everything else
(especially most code in xorn.geda) is refactored gEDA code.
> The only chance to get Xorn reception is to enhance the experience of
> users. Like fewer clicks for newbie-type users,
Given that the very idea of Xorn is to provide a better foundation, not to
add features, that's hard to deliver.
> like more options for experts, like solving problems which were
> unresolvable before.
That's more the idea of Xorn. Things which were difficult before and are
now realistically possible include
- proper scripting (this was one major incentive for me to start Xorn)
- solving the red dot problem
- allowing nets inside symbols (e.g. for schematic templates)
- disabling selected netlister stages (e.g. for cascade)
- passing more structured information to netlister backends (e.g. for
allowing subschematics to be routed as independent sublayouts)
- accessing all available schematic information from netlister backends
(e.g. for passing certain net attributes on to PCB)
> Seen in this light, Xorn is only 50% finished so far.
Much less than 50%. If you see the above list, this should be obvious. ;)
On Tue, 22 Sep 2015, Markus Hitter (mah AT jump-ing DOT de) [via
geda-user AT delorie DOT com] wrote:
> Xorn has a few new features, but also drops a lot of others.
The only feature Xorn drops is executing Scheme code, which is a direct
consequence of it not embedding Guile. This hasn't been an easy decision,
but the lack of proper Python bindings for Guile and the fact that Scheme
scripting would be kind of redundant tipped the scale.
- Raw text -