If someone really should be interested in the background of rubberband auto-routing of printed circuits boards, you can find more details in the PhD thesis of Tal Dayan from 1997:
[https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.8.6828&rep=rep1&type=pdf](https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.8.6828&rep=rep1&type=pdf) Unfortunately the links to his thesis sometimes changes, but searching with <https://scholar.google.de>/ with search term "tal dayan router" or "tal dayan 1997" should find it. Around 2010 we had discussed on the gEDA mailing lists concepts about autorouters for printed circuits boards. Beside the conventional grid based maze routers, which generates generally straight lines to connect electric terminals on a PCB board, we discussed topological routing with rubberbands. And I started an experimentation implementation in Ruby. Unfortunately gEDA became less and less popular and was mostly replaced by the KiCad project then, so it made not that much sense to try to integrate a rubberband router into the gEDA schematics editor and PCB layout tool, as number of users went close to zero. And it became clear that Ruby is much better than C suited for such a router, but is still not the optimal language. So the project was stalled for a long time. On the page <http://ssalewski.de/Router.html.en> there is a picture of a real PCB board with rubberband routed traces. The concept of routing copper traces on a PCB board is connecting terminal points with traces, while avoiding touching other terminals. Traces should be generally short, and of course are not allowed to cross. But most PCB boards have at least two layers for the traces, so traces can use vias to go from one layer to another layer. See <https://en.wikipedia.org/wiki/Printed_circuit_board> And a commercial project with rubberband routing exists also: <https://en.wikipedia.org/wiki/TopoR>
