On Mon, Feb 25, 2008 at 10:12:27AM -0500, DJ Delorie wrote:
>
> > > * All the new layers.
> >
> > Hard to parse.
>
> Not really; layer syntax is the same for all types. We just need some
> way of tagging which ones are which, and perhaps something in the UI
> to keep you from doing the wrong things on specific layers (like, for
> example, text on an outline layer)
>
> > But having a list of layers that describe the physical order of the
> > layers is necessary for blind and buried vias. On the other hand I
> > agree that the current layer grouping system has to be removed.
>
> I wonder if the layer grouping could *be* the stacking order? That
> lets us group silk, keepout, mask, etc, with the copper layers they
> correspond to, preserves the old usefulness of groups, and doesn't
> change the internals of PCB as much.
Ok, then I've no strong preference either way.
>
> > > * Support for footprint rotations at arbitrary angles.
> >
> > Yes, and this requires some file format changes.
>
> :FreeRotateBuffer(45)
>
> I think what you're asking for is for footprint definitions and
> footprint instantiations to be kept separate, so that rather than
> rotate the footprint, we just keep track of the angle separately.
> This is lossless and still allows us to update footprints
> automatically.
I'm not really in favor of automatic footprint update, at least you
should ask before modifying a board behind the user's back.
>
> I.e. a footprint instantiation should have:
>
> * ID of original footprint
> * position and rotation numbers
> * pin names (from gsch2pcb) and other attributes
> * possible overrides for things like pad size, shape, mask gaps, etc.
>
If you can essentially override everything, you are very close to
having the whole footprint in the .pcb file.
> > And while we are at it, why not switch to metric dimensions for
> > internal representation?
>
> Why? The only thing I'd recommend switching to is either (1) GCF[*]
> units, so we can describe both metric and english precisely, or (2)
> tagged units (i.e. the file says "2.1mil" instead of "210".
>
> [*] greatest common factor, so that neither unit is a non-whole
> multiple of it.
Well, GCF at least. When PCB had 1 mil resolution, GCF was 0.2µm
and it would have been reasonable to use 0.1µm. Now GCF is something
literally on the atomic scale (2nm), which may lead to very large
absolute coordinate values for large boards (I've heard of boards
several meters long with printed antennas on them but I've never
seen any). The point is that the inch has been defined as exactly
25.4mm for about 50 years and not the other way around, so GCF
will always be a fairly simple value in metric dimensions. I don't
really care if it is 1 or 2nm, although 2^31nm is "only" a bit above
2m while it is over 4m with 2nm step.
Regards,
Gabriel
>
>
> _______________________________________________
> geda-dev mailing list
> [email protected]
> http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev
_______________________________________________
geda-dev mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev
