[Question][libavoid] Experimental features under Avoid::RoutingParameter

[maxmrzk]

Hello,

I've been using libavoid for orthogonal connector Routing and like the results a lot. For a current routing problem, I would need to use some of the Routing Parameters which are flagged as experimental and not recommended for normal use. In detail:

• crossingPenalty
• clusterCrossingPenalty

I think these features add a lot for more complex graph structures and I would like to know, if there are any plans on making these non-experimental in the future. Or otherwise, if there are any other reasons besides performance that speak against using these parameters.

I just wanted to confirm that it is secure and reliable to invest time in developing on the foundations of these features. Thank you very much in advance !

[Aksem]

Hello,

I can share my experience:

• crossingPenalty works pretty well, but as you mentioned, it has a significant performance cost
• clusterCrossingPenalty in my previous tests, either I used it wrong, or it sometimes gave wrong results

I would also be interested in improving both. Also, I have a fork of adaptagrams, which I maintain.

[maxmrzk]

Thank you for sharing your thoughts and letting me know.