Ability to embed JavaScript into PDFs.
For large graphs, in my experience, the best way to view them is by rendering them as PDFs.
On the first sight, PDFs do not support any custom actions. However, at least in principle, we can embed JS into PDFs and execute that JS in PDF viewers that support that feature.
I'm wondering, is something like that possible today? That is, can we passthrough JavaScript code through .dot to generated PDFs and assign actions (e.g., clicking on some node runs some action) and have those actions be invoked in supported PDF viewers?