[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: pcb crooked traces



On Wed, Oct 13, 2010 at 12:48:14AM +0800, Steven Michalske wrote:
>
> We should have NO base unit internally, and only scale to internal
> units on import, export, open, and save.
>

I like this in principle, but it looks like it could get ugly
fast. Basically, we'd be creating floating-point units.
 
> The files should not use the internal representation of the units in
> the file format, and dosen't for low res footprints, ot ones defined
> with units.
> 
> The internal representation should be flexable.  If a board is sized
> to 200 meters, then pcb should adjust internally to allow for 428
> meters. Then pcb should set's it's internal units to 100nm and call it
> a day while warning that the minimimum resolution is 100nm.
> Or if you happen to be a 64bit pcb implementation it would be just
> another day at 1nm with that 428 meter board.
>

What happens when you open that 428m board with a 32-bit version
of pcb?

> If when you load a footprint and the resolution required to resolve
> the footprint is not there then it should warn or error that rounding
> is occuring.
> 

How do we know when there is rounding occuring? If I make a tiny
footprint, perhaps pcb will be working in picometers while I make
it --- but I only space things on a .5mm grid, so every measurement
is a multiple of 500 000 000pm.

When importing it, we'd have to analyse the footprint and see that
it could be loselessly scaled to a base unit of 10pm, 100pm, 1nm,
..., 500um, but not to anything greater.

If things are done in imperial (or worse), things are even more
confusing. Basically we would have to take the gcd of every single
coordinate of every footprint we import!


Compare this situation to one where everything is in nanometers,
except for a few oddball boards who are explicitly scaled to
something different.


Andrew



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