From 68f7e42c8a2f20b7568448fbe26b2cc0c9a47c83 Mon Sep 17 00:00:00 2001 From: Ethan Buchman Date: Sat, 7 Sep 2019 13:30:17 -0400 Subject: [PATCH] Apply suggestions from code review --- docs/architecture/tree-chunking.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/architecture/tree-chunking.md b/docs/architecture/tree-chunking.md index a7527a98a..bac699049 100644 --- a/docs/architecture/tree-chunking.md +++ b/docs/architecture/tree-chunking.md @@ -34,8 +34,8 @@ pairs alone. Such nodes would be part of the proof structure. The chunker can assign to each chunk a unique index, and every chunker will compute the same index for each chunk, assuming they follow the same strategy. However, it is not clear that, in general, a chunkee can verify the -chunk index, as there is no explicit map from chunk index to chunk or vice -versa. +chunk index, as there is no inherent map from chunk index to chunk or vice +versa. Such mappings may need to be provided externally and trusted, like in the case of Warp-Sync. This raises two concerns for a state sync protocol: - how to handle mis-indexed chunks (ie. requesting chunk 3 and receiving chunk