Re: gEDA-user: A couple o' questions

[John Sheahan <jrsheahan@xxxxxxxxxxxxxxxx>]

It becomes a real can of worms - but if you have a
design specific library, even an _instance_ specific library,
you can get apply attributes that function.

Thats very unfriendly to a beginner however.
And only necessary for analog design.  Or fpga pins.
(or a very long list when I consider for a moment..)

john




John Doty wrote:
> On Apr 7, 2006, at 5:43 PM, John Sheahan wrote:
>
> > Mark Rages wrote:
> >
> > > The pintype *is* obvious for the "fistful of TTL chips" approach to
> > > digital design, which was commonplace when EDA systems were first
> > > getting developed.  Maybe the pintype DRC is a bit of an anachronism.
> >
> > maybe DRC applies much better to some types of designs  than others.
>
> Yes!
>
> > Which would imply to me that DRC should be used judiciously.
>
> Yes! That's the problem with gnetlist's drc2 hard sell.
>
> > Perhaps the components themselves should know whether that are good at 
> > being DRC'd.    Or a DRC rule decks be chosen for purpose?
>
> I think that having pintype default to "pas" would be constructive. That 
> should essentially turn off the pintype DRC for a component whose author 
> couldn't figure it out.
>
> > I never figured out the DRC attributes for a bipolar tansistor at 
> > board level for exmaple.  john
>
> Yep. The roles of the pins aren't properties of the device: they're 
> determined by how the device is used. The emitter might be 'in' (common 
> base), 'out' (common collector), or 'pwr' (common emitter). So, use the 
> escape hatch and call it 'pas'. Won't get checked, but won't cause an 
> error. Yeah, I know, that's not obvious unless you understand the 
> details of how DRC works (or doesn't). That's why 'pas' should be the 
> default.
>
> John Doty              Noqsi Aerospace, Ltd.
> jpd@xxxxxxxxxxxxx