← Back to team overview

kicad-developers team mailing list archive

Re: Atomic Libraries Proposal



just a possibly stupid idea. Maybe it could be an option to move part of the ideas of the new library system into your third entity. Particularly the inheritance stuff might not really be required if your atomic entity is used.

I seem to remember there was talk that this would be quite hard to implement anyways. So if that would make implementation easier then this might be a good way to go. (But maybe i simply missed the point of why inheritance or your system is useful as i kind of thought the inheritance stuff was in there to fulfill exactly what your third file would now do.)


You state in the format definition that you do not intent to change anything regarding 3d models. Maybe your file could really go a bit further than you first envisioned and could take care of all possible connections in one system. So lets say it includes not only symbol footprint and bom information but also an optional overwrite for the 3d model as well as a way to define spice models. And possibly also a way to define the symbol pin to footprint and spice model pin mappings. (I would assume that this might be a bit too far so feel free to reject this outright as i must admit i have not yet fully convinced even myself that this would be a good idea. This could even be something that is moved to a second iteration.)

On 24/05/19 18:25, Seth Hillbrand wrote:
Hi Rene-

These are good points, thanks for helping me clean this proposal.

One design goal of this system was to make it transparent for both
librarians and users who do not want to use it.  To that end, the
existing design of both the current and new formats is not affected.
Users who do not enable this option would see no change in their

This does mean that there are potentially two different places where the
same key can be defined.  As envisioned, no fields from the symbol
library will be used when an atomic part is placed on the schematic.
All properties from the symbol library are cleared and the atomic
library fields are populated.

I will clarify this in the usage section.


Am 2019-05-24 09:42, schrieb Rene Pöschl:
Hi again,

A few clarification questions:

Would your proposed solution replace the way of defining footprints in
the symbol file format of the future or would this be in parallel to
it. (I lean towards the fact that it should replace it as i fear we
otherwise end up with something similar to what we had with the
datasheet being stored in either dcm or lib files. But i can also
imagine that this would not be very well liked by the community)
Similarly with footprint filters. In the official lib we use it for
fully specified symbols to allow adding manufacturing specific
variations. (Inclusion of thermal vias, hand soldering alternative,
...) How would that come into play?

Your system would take care of adding bom info via properties. Right
now such fields can already be added to a symbol directly or via
eeschema. Again something where we might want to either limit it to
one place in the library or have clear rules which place takes
precedence. (An interesting case is if there are different fields in
the symbol compared to the new part specifier. Should the result be a
union of both sets of properties or should only the part properties be

On 23/05/19 21:41, Seth Hillbrand wrote:
Hi All-

After some discussion, we are trying to decide whether implementing a
basic atomic library support would be useful during v6 or would cause
more headache than worth while trying to work with the solution.

Background: "Atomic" libraries are ones that have unique names for
symbol+footprint+properties combinations.  These are also referred to
"Fully-specified" libraries or sometimes "house libraries".

The paradigm is outlined in a google document[1].  This is meant to
provide an explicit option for users to create/utilize atomic library
components without affecting the symbol or footprint formats.  It is
also meant to be zero impact to the current, official KiCad libraries.

The major limitations of the specification are:
- No editing of the atomic library file within KiCad
- The atomic library does not contain symbol or footprint data

Folks are welcomed to weigh in on whether this approach presents a
viable work flow for them.  I'm happy to take suggestions on the
approach, I may ask you offline for some additional details of how you
use KiCad and/or other tools with libraries at the moment.



Mailing list: https://launchpad.net/~kicad-developers
Post to     : kicad-developers@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~kicad-developers
More help   : https://help.launchpad.net/ListHelp

Mailing list: https://launchpad.net/~kicad-developers
Post to     : kicad-developers@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~kicad-developers
More help   : https://help.launchpad.net/ListHelp