Currently there are many edges, which IMHO makes the graph hard to read.
OTOH, @smarr expressed it's better to have all the information, and then filter when presenting.
Some things I'd like to simplify based on the current graph:
Concepts edges seem mostly useful for the "root" paper, I think it's OK to assume a paper extending another one "inherits" the concept and not show that edge explicitly. e.g. Lookup Caches.
My initial use of concepts was purely to have something to link to if there was no parent paper. So maybe we should have a way to only show concepts which have a paper extending it.
Concepts seem very much like tags, each paper could have many of them and if we wrote them it would for sure be unreadable. I'm thinking concepts should only be shown in the preview below, and not in the graph?
Tracking Reachability to Reduce Synchronization is not really a concept, more like an idea connecting multiple papers using that idea for various purposes.
Currently there are many edges, which IMHO makes the graph hard to read. OTOH, @smarr expressed it's better to have all the information, and then filter when presenting.
Some things I'd like to simplify based on the current graph:
Lookup Caches
.extending
it.Tracking Reachability to Reduce Synchronization
is not really a concept, more like an idea connecting multiple papers using that idea for various purposes.