pull down to refresh

This Week in Bitcoin: Bringing the “NOW” to the Orange Dev Suite
2198 sats \ 13 comments \ @sorukumar 16 Jun bitcoin

All,

Here is a new addition to the orange-dev-suite: This Week in Bitcoin.

It surfaces PRs merged, hot PRs, hot discussions from Delving and the mailing list, plus contributor highlights. Each with short, pleb-accessible (value-focused) AI summaries.

I may schedule it as a weekly post on X and LinkedIn also next week after incorporating feedback. Depending on how it goes, we may have a human review/edit the AI write up for each PR and discussion.

On the “why”:

I’ve been exploring ways to create a roadmap of sorts for Bitcoin. h/t @schmidty (X Post). This felt like a practical first step using the same data pipeline behind Orange Dev Tracker and Network

The broader suite now gives a clearer picture:

I’ve cleaned up the dashboards for better consistency. Still very much WIP. Feedback tab is live (a feedback from @koob ) .

Pass on your thoughts on what works and what can be improved.

write
compose
related posts
113 sats \ 2 replies \ @Murch 18 Jun

How are "Retired", "Fading", and "Steady" determined?

E.g., in this screenshot, Christian Decker is marked as retired, but actively working on CLN,

Nicolas Dorier is marked retired, but maintainer of BTCPay Server and NBitcoin,

I haven’t really seen anything from Aaron Voisine in forever (had a hard time reaching him regarding some of his BIPs), while he’s marked as "Fading",

and Brandon Black is marked as retired, but has contributed review to the BIPs repository (and Optech) this year.

Just saw, that Steven Roose is also marked as retired, but he co-authored a BIP this year and is working on one of the Ark projects.

reply
22 sats \ 1 reply \ @sorukumar OP 19 Jun

I'll have a look and report back.

Basically, there's super low activity in the last 3 years compared to their past contributions. I ended up tweaking and reached to the definition to mark Gavin Andresen as retired . he was showing up as active because of his 2022 comment saying he is retired.

At a high level, if someone had, say, 20 reviews before the last 3 years and only 1 in the recent period, they're labeled retired. less than 5 reviews makes them "fading."

I haven't accounted for work in btcpayserver or NBitcoin or Lightning yet, so I see that as a data limitation. However, I need to rethink the methodology , it's even marking 2026 contributors as retired.

reply
1 sat \ 0 replies \ @Murch 19 Jun

Yeah, I think Gavin left the space in 2016, he should definitely show up as retired

reply
1 sat \ 8 replies \ @Murch 16 Jun

Thanks, that’s cool! One nit: I noticed that you list Sebastian van Staa as a new contributor (svanstaa), but I think he has had several PRs merged in Bitcoin Core (around 5 or 6?), there might be an issue there.

reply
132 sats \ 4 replies \ @sorukumar OP 17 Jun

Great catch. You're right. Sebastian has dozens of PRs merged in bitcoin-core/guix.sigs.

However, our current data pipeline only tracks the main bitcoin/bitcoin repository. Since PR #34636 was his first PR merged into the main repo (on June 10th), the script flagged him as a new contributor this week.

To make this clearer on developer profiles, I'm going to start displaying both the authored commit date and the PR merge date.

Quick question: As we expand our data sources (we're already adding BIPs), do you think we should also include satellite repos like secp256k1 and bitcoin-core/gui in our main analytics?

reply
41 sats \ 3 replies \ @Murch 17 Jun

Sorry, I made a mistake with my search parameters regarding svanstaa, I agree it’s his first PR he authored.

Yeah, secp256k1 and bitcoin-core/gui would make sense to include.

reply
1 sat \ 2 replies \ @sorukumar OP 18 Jun

Thanks for the feedback, @Murch. That's a big reason we've been able to make all these improvements to the dashboard.

added secp256k1 and bitcoin-core/gui, plus repos like bitcoin-core/guix.sigs. on the profile page, though, I still break the commits into 'core' and 'ecosystem'.

here is how svanstaa profile looks like now.

reply
1 sat \ 1 reply \ @Murch 18 Jun

Interesting idea! Glancing at this, the same separation should then maybe also be applied to the other metrics then. E.g., PRs Reviewed.

reply
22 sats \ 0 replies \ @sorukumar OP 19 Jun

Will add it in the next iteration.

reply
114 sats \ 2 replies \ @Murch 16 Jun

Looking around on the website a bit, I noticed that the list of maintainers seems to have some quirks. Here are a few things I noticed:

  • Luke Dashjr, Cory Fields, and Carl Dong were never Bitcoin Core maintainers.
  • Sebastian Kung only became a maintainer this year
  • Ava Chow has been a maintainer since 2021
  • Marco Falke was a maintainer from 2016–2023, not 2015–present
  • AFAIK, Jonas Schnelli started in 2015, not 2013
  • Gavin Andresen was maintainer until 2016, not 2015

This list should be accurate and also has sources for most dates: https://bitcoin.stackexchange.com/a/88649/5406

reply
22 sats \ 1 reply \ @sorukumar OP 17 Jun

here are the details:

I cross-referenced the StackExchange link you mentioned. Our dataset[https://github.com/sorukumar/orange-dev-data/blob/main/metadata/maintainers.json] actually matches the StackExchange details perfectly, but I had supplemented it with commit data directly from the repository, which caused some quirks in the UI.

  1. On the dates for Gavin and Jonas: The data pipeline was dynamically inferring their timelines based on their earliest and latest merge commit dates rather than their official appointment dates. In Gavin's case, his last actual merge action was in 2015 (even though he held keys until 2016). For Jonas, he had made a local merge commit on a PR branch back in 2013, which tricked the script into expanding his timeline. I've updated the logic to strictly use their official appointment dates, so the chart is now historically accurate.
  2. On Luke Dashjr: He does have about 9 merge commits in our data from the 2011-2016 era. I assume this was before the modern trusted-keys protocol for merge commits was strictly enforced.
  3. On Strategic Maintainers: In the UI, I use dotted/hollow bars for Cory Fields, Carl Dong, and Sebastian Kung (sedited) from 2019–2025. This visually distinguishes that they were Strategic Maintainers (e.g., maintaining the build system, security, or Guix) but did not hold keys to merge into the master branch.

Refining the UI for this chart is still a pending item. I'm planning to move it from the Tracker ("what") over to the Network dashboard ("who") for a cleaner separation of concerns in next iteration.

By the way, what do you think about the approach of visually distinguishing and showing Strategic Maintainers alongside the traditional Committers?

reply
95 sats \ 0 replies \ @Murch 17 Jun

Regarding Luke’s merge commits, it looks like the same thing happened there that also happened to Jonas. I tracked down there of his merge commits. One was the last commit in the PR https://github.com/bitcoin/bitcoin/pull/7192, another was the second in https://github.com/bitcoin/bitcoin/pull/7192, the third was the fourth in https://github.com/bitcoin/bitcoin/pull/505. The first two PRs later got merged by Mara van der Laan, the third by Gavin Andresen.

On Strategic Maintainers: In the UI, I use dotted/hollow bars for Cory Fields, Carl Dong, and Sebastian Kung (sedited) from 2019–2025. This visually distinguishes that they were Strategic Maintainers (e.g., maintaining the build system, security, or Guix) but did not hold keys to merge into the master branch.

At least among Bitcoin Core contributors the term maintainer has a very specific meaning. I’m aware that these three led big projects, but I haven’t heard the term maintainer in that context before. Maybe a different term such as "project lead" would fit better?

reply
1 sat \ 0 replies \ @nitter 16 Jun
reply