geda-developers team mailing list archive

[DJ Delorie <dj@xxxxxxxxxxx>]

> Irrespective of what tools would be written or changed what do you
> imagine the user experience of this to be?

I think I grand-schemed a while back, don't recall any more.  I see a
few avenues of work that can be explored, although I have no idea how
they'd be implemented and what they'd look like.

* PCB's internals need updating still.  The layer flags is a good
  start, but we really need things like blind and buried vias,
  subcircuits (corresponding to gschem's heirarchy, for example),
  microstrips, matched pairs, push-n-shove manual routing, better
  footprints, and better autorouting.  This is stuff that is "normal"
  on commercial layout tools and users ask for these a lot.

* We need something to solve the component problems.  The transistor
  problem, light vs heavy, layout vs sim, etc.  I've blue-skied some
  of this but that's just my ideas.

* We need better base libraries.  Hopefully we can just adopt
  KaiMartin's, since that's the whole point of his library.  That just
  leaves the "what about breaking old schematics?" problem to solve.
  We need to be able to import/reference more than one library at a
  time, too.

* More workflows need to be able to be done in a gui-only environment,
  targetting common projects and new users.  A prime example of this
  is being able to go from schematic to simulation with a push of a
  button.  PCB's File->Import is an example of this.  We're used to
  working off the command line with scripts and Makefiles but new
  users will be expecting the tools to be usable in a desktop
  environment too.

* Back/Forward annotation is still a hot button for users.  I haven't
  looked at Igor2's work yet (it's on my list ;) and I'm hopeful for
  it.  But consider annotating schematics with simulation results,
  importing simulation results into pcb layouts ("this test point
  should look like this waveform") etc.  These are all examples of
  project X sending data to project Y outside the "normal" flow.

So looking at it from an "integration" point of view, there are three
problems:

1. Managing components and libraries across tools.

2. Making it easier for users to combine tools into workflows.

3. Making it easier for tools to share information.