Sender: Richard Barlow <skkinny AT gmail DOT com>
Message-ID: <1355863384.24123.93.camel@thinkpad.richardbarlow.co.uk>
Subject: Define exactly what the DRC does (was: Re: [geda-user] Find rat
 lines)
From: Richard Barlow <richard AT richardbarlow DOT co DOT uk>
To: geda-user AT delorie DOT com
Date: Tue, 18 Dec 2012 20:43:04 +0000
Content-Type: multipart/signed; micalg="pgp-sha1"; protocol="application/pgp-signature";
	boundary="=-JGZPo7kY9wOCptQz/lBP"
Mime-Version: 1.0
Reply-To: geda-user AT delorie DOT com


--=-JGZPo7kY9wOCptQz/lBP
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tue, 2012-12-18 at 19:56 +0000, Peter Clifton wrote:
On Mon, 2012-12-17 at 14:13 +0100, Markus Hitter wrote:
>> Am 17.12.2012 um 10:18 schrieb Levente:
>>=20
>>> Take the situation when you have A_GND and D_GND net for analogue
>>> and digital ground. You WANT to short it in one point. I usually do
>>> it with a 0
>>> Ohm resistor, or I short the nets in PCB.
>>=20
>> You could make a footprint with just line in it, representing the =20
>> bridge between both nets. On the manufactured board you wouldn't =20
>> see it at all.
>=20
> We should probably be able to track that as a short too.. or a DRC
> fail / warning.

I've never seen it explicitly stated anywhere, but from my experience
I've built up the following definitions/distinctions for the PCB DRC and
ratsnest optimisation:

All the DRC does is guarantee that the board is reliably manufacturable,
it doesn't care about the electrical connectivity being correct. If the
board passes the DRC then a manufactured board should (if the DRC
doesn't have any bugs) match the designed layout.

When optimising the ratsnest, PCB checks the electrical connectivity and
informs you if there are un-routed connections remaining or shorts
between nets.

I'm not saying that the way this is split is necessarily good - I know
of at least one person who has been caught out assuming the DRC would
inform them about shorts - but it is currently quite cleanly split. If
you are considering bringing up connectivity errors as part of the DRC
process then all connectivity errors should be in the DRC (such as
un-routed connections) if not then the the definition of what the DRC
does and does not do should be made more clear.

Personally I'm in two minds with this. Technically what the DRC does
currently is correct, it ensures the design follows the design rules.
However from a usability point of view, not having a single central
place to provide a list of problems with the design is bad.

Rich

--=-JGZPo7kY9wOCptQz/lBP
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: This is a digitally signed message part
Content-Transfer-Encoding: 7bit

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)

iQEcBAABAgAGBQJQ0NVYAAoJEMWwul/B8SCJMgQH/1TDQtrjomjmAP58DxFv9+nf
HXxBTkIlSfdTYMB0xaHbBQqpKQNYObhsPVtG+H+z91O/AgN+Raxehz/KY7aEU0MV
i2dm+6xhV2O656nV6v4SHs0Ne/Wjsx/XzNfZl7a8/Lj18omVMfP+YvVUYMpcO3nt
dtcrXHp0VAK2k4nGVMD+GaORd3XAOGa20dzrROAxBlTzhQOOV2qd7vG3xAYYCFuG
MtTogpeo9orhWQe45KGuv+1CbsnAskn+sVEC+niQ5Qw1ZLnABQbwcuv+96jpNji2
d6AEK+sK8/I8oip1ioq9wSTaXNNQbmH6eR//TDmn3meZdLjpe1xWjMOV9HxYqjQ=
=oDa+
-----END PGP SIGNATURE-----

--=-JGZPo7kY9wOCptQz/lBP--