aboutsummaryrefslogtreecommitdiffstats
path: root/docs
diff options
context:
space:
mode:
authortv <tv@krebsco.de>2026-11-06 04:30:45 +0100
committertv <tv@krebsco.de>2026-11-06 04:30:45 +0100
commita56afd335558eda0fee6f511cc9469d3848e5637 (patch)
tree5a3cd6c933de18da6bb270e6eddead5a866e2f01 /docs
parent278ca3cd21397ffad07f0874b80b77f41088a3f2 (diff)
views documented
Diffstat (limited to 'docs')
-rw-r--r--docs/peers.md11
1 files changed, 11 insertions, 0 deletions
diff --git a/docs/peers.md b/docs/peers.md
index 2ff10e0..13ee85d 100644
--- a/docs/peers.md
+++ b/docs/peers.md
@@ -380,6 +380,17 @@ reaches no other member, by its parent's), or can't apply the next message and a
what it lacks, and gets just that, and what that builds on, again (not what was sent
after it: that may only be late).
+The parent and the child each work the tree out for themselves, and they agree only as
+far as they have the same inputs. The member list's graph and `k` are in the document;
+whom a member reaches and its round trips, it tells its neighbours (its parent among
+them); but whether a member a hop closer reaches the writer's side only its own
+neighbours know, and a node that doesn't takes it as reaching it. Where they differ, a
+change goes twice, or the parent counts the child as covered by another while the child
+counts on it: then nobody sends, and the slow round brings the change. So does it where
+the parent doesn't reach a member that reaches it and picks it (one-way reachability):
+the parent sends it nothing, and counts nothing. The benchmark counts both (child links
+the child confirms, and orphans: `simtest/bench.rs`, "views of the trees").
+
A node's neighbours that get a writer's changes from another parent in its tree, and
the writer, count as having them: what it passes on to them later along other trees
doesn't carry them again. Without that, with writers in turn, each change coming along