To use via/pin/pad from a predefined library and be a= ble to update them would be very nice then a small adjustment is needed.

Sometimes a pad need to have a little bit unusual sh= ape and in such cases it is very useful with possibility to add the necessa= ry shape on the needed layer. For example a power pad or antenna.

I am looking at clearance and maybe it is possible with= a small step in this direction at the same time.

= Nicklas Karlsson

<= div class=3D"gmail_quote">2016-01-28 21:46 GMT+01:00 Britton Kerin (britton DOT kerin AT gmail DOT com) [via geda-user AT delorie DOT com] <geda-u= ser AT delorie DOT com>:

On Thu, Jan 28, 2016 at 11:08 AM, Stephan B=C3=B6ttcher <geda AT psjt DOT org> wrote:
> "Britton Kerin (britto= n DOT kerin AT gmail DOT com) [via geda-u= ser AT delorie DOT com]"
> <geda-user AT delorie DOT com= > writes:
>
>> It's highly unlikely it'll be easier for just BBVias.=C2= =A0 I believe
>> there's already an experimental branch that implements them.= =C2=A0 Igor is
>> likely to decide to implement them again while we all talk about i= t.
>> For containers to pay off several other applications for them woul= d
>> need to exist and be important enough to get implemented.=C2=A0 Wh= at are
>> they?
>
> My use cases for a mor egeneralized data model:
>
>=C2=A0 - pad stacks, with differnet pad sizes on middle layers

Everyone seems to want these, to be honest I don't understand wh= at
pads on middle layers are for.=C2=A0 Not mounting components for sure.=C2= =A0 If
it's just a node what's wrong with a trace or poly?

>=C2=A0 - hierarchical layout

It can be done now...

>=C2=A0 - rigid-flex baords with more than two outer layers

I've never done these.=C2=A0 I guess the flexing connector layer= is on top,
but only where it exists so you can still have components on the layer
below?
There's still only one top layer at any given point?

> To avoid more spagetti special casing in the future for all kinds of > use cases.

I've just been through a lot of DRC and while it's true that= e.g.
geometrical tests could be concentrated in Union types or something,
they're still gonna have to go somewhere and it wasn't that bad to<= br> find them as they are.=C2=A0 Adding abstractions that aren't universall= y
utilized doesn't necessarily make a code any more comprehensible,
because sometimes-utilized abstractions are arguably more confusing
than none at all.=C2=A0 To justify things like containers on
code-cleanliness grounds a big refactor effort is implied.

Britton