X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f X-Recipient: geda-user AT delorie DOT com Message-ID: <1445006837.2017.17.camel@ssalewski.de> Subject: [geda-user] Is someone working on a PCB autoplacer? From: Stefan Salewski To: geda-user AT delorie DOT com Date: Fri, 16 Oct 2015 16:47:17 +0200 Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.16.5 Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Reply-To: geda-user AT delorie DOT com Maybe I missed that, but there was so much traffic on this list in the last months that I was able to read only a very small fraction. An autoplacer has the advantage, that it may be useful also for people who wants to do manual routing. And for autorouting it is an important component. Another point is, that even a bad autoplacer is useful for initial part distribution, so all the arguments people have against autorouter are not valid for autoplacers. And development may still be some fun. For an autoplacer some support by schematics tool would be really useful, for example grouping of elements: An OpAmp with its two resistors and two bypass capacitors should build a group, indicating that these components belongs together. This is important for the bypass capacitors, because their pins both belong to very large nets, gnd and power, so location is not determined by net. For signal paths, nets should include generally only two or very few pins/pads, so that placer can use that information to place components at the correct initial position. Another point is, that it would be useful when footprint files would contain a bounding box for occupied area -- pads, pins and silk may be used for that, but exact box may be better, maybe not only a rectangular box, but a polygon or circle.