Re: gEDA-user: General Layers questions

[DJ Delorie <dj@xxxxxxxxxxx>]

> Current state:
> One physicaly connected surface is called "layer group".
> What is called "layer" is something that one can work on
> and will end up in some surface "layer group".
> There are special "Layers" called silk, ratlines, mask, outline,..
> 
> Proposed state:
> I think it is common to call "layer" one electricaly
> connected sufrace. Cannot we come up with some other
> name for "layer", so we can leave "layer" for one electricaly
> connected surface?

We've discussed this before, came up with something like "drawing
sheet" vs "physical layer".

> If there are more (ie. silk) layers, where to put the silk
> layer of an footprint? Should we somehow link cooper
> layer and silk layer? What if we link more than one silk
> layer to cooper layer?

Aside from numbering layers from 1 .. N, there are symbolic names for
"top", "inner", and "bottom" layers.  Footprints use symbols names,
they're mapped to numeric layers as appropriate for each board.

That doesn't mean footprints *can't* use numeric layers, but they're
more reusable if they don't.

Each drawing layer has a record of which physical layer it corresponds
to (if at all), so you could (for example) assign a silk layer to one
of your internal layers, despite that not making sense.  However,
assigning a documentation layer to an inner layer *might* make sense.

> Or do we want to have some "default" silk and cooper layer and move
> all footprints from that one to desired one?

The stackup needs to understand the concept of inner and outer layers
at least, so it can map a library footprint to a specific element.  Of
course, I leave open what "outer layer" means, esp in the context of
flex cables where a given layer might be an inner layer over here, and
an outer layer over there...

> One of the problems someone wants to solve is to be able to put
> "outline" of an footprint to one silk layer and "refdef" to another.

Again, we need to have some way of mapping a well-known set of
symbolic names to specific physical layers.  For text, perhaps we
adopt gschem's named attribute method?  So we have some text *named*
refdes, with *value* of "R4", that happens to be on layer
"topsilk:refdes" - it looks for a "topsilk:refdes" mapping rule, and
if it doesn't find one, looks for a "topsilk" mapping rule, etc.



_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user