kicad-developers team mailing list archive
Mailing list archive
Re: Internal PCB Units?
--- In kicad-devel@xxxxxxxxxxxxxxx, "vladimir_uryvaev" wrote:
> Main reason is exactness, not exceeding precision. As i wrote
> before, rounding sometimes makes bad thing. Values like 0.1992 mm
> may fail DRC at manufacturer side causing board not to be produced.
> This is why quantum should be metric: every inch value may be
> represented exactly in metric units, but not vise versa.
I *fully* agree that the quantum distance should be metric (rather than imperial), and as such, I would suggest either 12,700,000 or 25,400,000 quantum units to the inch.
Selecting such a distance would (also) permit imperial distances which are integral multiples of 0.01 mil (10^-5 inch) to be specified *exactly* (either 127 or 254 quantum units), while distances of a genuinely metric nature (e.g. 0.8mm) would always be "fully preserved" in the event of users switching between imperial and metric measurement units.
I speak from experience with using Protel and Altium Designer. A quantum distance of (1*) 10^-7 inches has been used with those applications for some time now, and genuinely metric distances have a pronounced tendency to be "mutilated" as a consequence of that choice of quantum distance...