[net2-wg] Should CTP header have hop count?

[Rodrigo Fonseca]

> I agree that it may be useful. The case would be stronger if it is
> known to be useful. You are saying this is useful for protocols like
> BVR?

Well, BVR's coordinate system is *built on* hopcounts, but it is not
clear that we would use the same tree building code, as BVR needs to
know the distance to all roots, and not only to the closest. S4 uses
distances. Every node is the root of a tree, but the trees are scoped,
and never deeper than the closest beacon. PathDCS also builds multiple
trees, and uses distances in hops to route.

A quick grep on the tinyos-1.x tree for "call RouteControl.getDepth()
shows that files in these directories

beta/SystemCore/MementoMori
beta/TinierDB/TestTimeSync/
contrib/GGB/apps/zmgent/
contrib/GGB/tos/lib/Straw/
contrib/ucbRobo/apps/sensornet/MagMHopRpt/
beta/TinierDB/TestTimeSync/
tos/lib/VM/languages/motlle/matelib/
tos/lib/VM/components/

used the call.
Straw, for example, used the depth in the tree to control the delay
between successive sent packets, before Flush offered dynamic rate
control.

Rodrigo