On Sep 4, 2020, at 9:11 AM, Roland Lutz <rlutz AT hedmen DOT org> = wrote:

On Thu, 3 Sep 2020, John Doty wrote:
On Sep 3, 2020, at 6:01 PM, Roland Lutz <rlutz AT hedmen DOT org> = wrote:
> On Wed, 2 Sep 2020, John Doty wrote:
> > I favor Lepton [because] it=E2=80=99s agnostic = about which layout program > > you use: it has no favored layout = component. [=E2=80=A6]
> > The layout people I work = with all have different favorite tools, so > > decoupling = schematic entry and layout is very important to me.
> = > This is absolutely the same for gEDA/gaf.
No, = gEDA/gaf is joined to PCB. The tail wags the dog.

gEDA/gaf supports a variety of = layout programs via netlister backends, some of which you contributed = yourself. There's even backends for SPICE and cascade simulation = which doesn't involve layout at all.

Yes, but = there=E2=80=99s no new development happening here. I was the last = holdout. There seems to be neither developer nor user interest in = anything except PCB support.

Xorn is an internal component; you don't have = to understand or even consciously interact with it in any way. = (But even if you did, it's not more complicated than the Scheme = API.)
Except if you=E2=80=99re writing back = ends, you do.

There's a Scheme = and a Python variant of the gnetlist API. You can use either to = write a backend.

I=E2=80=99ve done things in Python. It wasn=E2=80= =99t the problem. The problem was to figure out the API.

Let's stay at the example of the = Osmond PCB backend. It accesses the netlist in the following = ways:

- netlist.packages (the packages in = the netlist)
- netlist.nets (the nets in the netlist)

- package.refdes (the refdes of the = package)
- package.get_attribute('footprint', 'unknown') = (gets the "footprint" attribute of the package, with "unknown" as the = default)

- net.name (the name of the = net)
- net.connections (a list of pins connected to the = net)

- pin.number (the pinnumber)
- pin.package (the package to which the pin belongs)

What about that do you consider hard to figure = out?

When I first looked at this years ago, there was = no API documentation, just a very confusing high level description of = Xorn. Maybe there is better documentation now, but I=E2=80=99ve moved = on. And as far as I can tell, nobody but you has actually used this API = to write anything. It seems to be your personal environment in which you = are impressively productive but nobody else is. At least the old gEDA = scheme API, poorly documented as it was, attracted = contributors.