Two weeks back I shared orange-dev-tracker, a little project analyzing the Bitcoin Core repo. But as I started layering in mailing list data and Delving Bitcoin conversations, something became really clear: the social layer is a beast. Deserves some respect, and its own repo.

So I built orange-dev-network — a living map of 15+ years of technical consensus formation in Bitcoin

Here's the thing: repo commits tell you what got built. But the social layer? That's where you learn why — and more importantly, who actually convinced the network to move

📊 5 Key Insights from the Network📊 5 Key Insights from the Network

The BIP Paradox
My favorite data point. Only 14.6% of technical threads actually mention an explicit BIP number. This highlights a crucial reality: the BIP process is a tool for documentation, not discovery.

Mempool & Fees (2024 Peak)
Discourse reached a historical peak in 2024. The data shows a shift from general fee estimation toward high-complexity topics like cluster mempool, as the network adapted to unprecedented congestion levels.

The Covenants Explosion (2025)
Currently the most explosive theme in the data. While research has been constant since 2022, activity doubled in 2025, marking a clear transition from abstract research into a "concrete review" phase.

Lightning's Second Wave
LN technical discourse reached an all-time high in 2025. The data suggests a secondary "R&D wave" that is significantly larger than the initial post-SegWit period (2018), with a modern focus on protocol safety and channel jamming mitigations.

The Nic Carter "Influencer List" Reality Check
Considering there are not too many devs, it is hard to get it 'wrong' in any list.

However, here is what social data shows.

🛠️ Methodology: How We Actually Map Influence🛠️ Methodology: How We Actually Map Influence

Data Sources:

Bitcoin-dev Mailing List: Ingested from the lore.kernel.org/bitcoindev Public-Inbox git shards.
Delving Bitcoin: Aggregated via the jamesob/delving-bitcoin-archive repository.
Bitcoin Improvement Proposals (BIPs): Tracked via the bitcoin/bips repository.

The Algorithm:

Timeframe: June 2011 – February 2026 (27,800+ messages parsed).
Influence Calculation: We use Temporal PageRank (a time-aware version of Eigenvector Centrality). We treat replies as "technical peer review." When a high-influence contributor (like a Core Maintainer) engages with your thread, your own centrality increases. It’s a digital reputation system—it's not about how much you talk, but who listens and responds to you.

🙏 Need your Feedback🙏 Need your Feedback

@Murch , pls have a look, and let me know if something looks off.

cc Pleblab friends: @koob @Car

• New Views & Data: Any new cuts or datasets we should surface that we're currently missing?
• Pre-2011 Data: My dataset starts June 2011. If anyone has a clean, parseable source for the 2009-2011 SourceForge mailing list archives, please let me know. That gap bugs me.
• Data quality: Does our Influence Map miss anyone who's a "quiet giant" in the technical archives? Someone whose impact is real but maybe not visible in reply graphs?
• Identity Mapping: Check the interactive map — does my deduplication of emails feel accurate? I want this to reflect reality to the extent possible.

Check out the map: orange-dev-network

Previous work: orange-dev-tracker

P.S.: If you spot yourself in the network and the data looks wrong, pls respond. I'd rather fix it than be "technically" accurate about the wrong thing.