Wasabi Wallet (a popular Bitcoin Wallet) is currently running a Wasabi Contribution Game, where contributors to the Wasabi Wallet repo will be rewarded 1 BTC ($11,511 at time of writing) according to their contributions (merged PRs) over a one month time period, which we’re in the middle of ( 06-25-2019 to 07-25-2019). The merged PRs are scored by total lines added/deleted, a fairly crude metric. SourceCred can surely do better. So last week I ran SC on their repo, then used a script @decentralion created to host the instance on GitHub (chart).
Here’s the timeline view of their repo. NOTE: the contest had just started at this point, and SC hadn’t yet picked up on new contributions for the contest.
I Tweeted the chart to Wasabi Wallet, hoping to the maintainer’s attention.
Didn’t get a response (yet), but got retweeted to @wasabiwallet’s (7,154) followers and was invited to their Telegram channel where someone reposted the tweet.
Today (7/6), since the contest has been running for 13 days now, I re-ran timeline SC on the repo. Didn’t see the new data show up for some reason, so filed an Issue about it on the SC GitHub. Was able to see the new data by just creating a fresh repo and running SC again (link). Below is the same chart, but with the first week of data during the contribution game.
Indeed, we can see a nice big spike! Looks like the game is working (or people are working the game;).
Plan to check back in when the contest ends, compare the official results to the SC scores, and try to engage the maintainer and/or contributors for feedback.
Neat! Maybe tweet this result at them now? I’m sure they’d be interested to see the quantitative (and dramatically visible) impact of their contest.
An idea for analysis: maybe set the node weights to assign 0 weight to pull requests. This would give a sense of how the “complimentary” activity (comments, reviews, issues) is growing alongside the pulls. Could be interesting to look at, since they are only strictly incentivizing pull requests.
Just re-ran SC to get the latest activity (timeline).
Below are screenshots showing the differences in cred for the most recent week, comparing the default (PR weight = 4) to PR weight =0.
PR weight = 0:
PR weight = 4 (default)
At first glance, it appears that setting the weight of PRs to 0 mainly just gives more cred to the maintainer and another core contributor. Which is expected, as they’re probably busy reviewing PRs of (potentially new) contributors? Seeing anything more interesting in the data?
I think this is a good opportunity to explore what kinds of questions an actual user of SC wants to answer. In this case, I assume it’s primarily to recruit new people. Though they could also be experimenting with performance pay generally; the Bitcoin community in general does have a long history of experimenting with new ways to fund development.
Tweeted this to them.
Just checked the views on the two Tweets I posted about this. Views have apparently roughy doubled, to 6k and 7k, respectively. Don’t know if that’s good or bad for crypto Twitter, as I’m new to it.
Totally! Do we know of other projects that might be interested in exploring SourceCred?
Also, @s_ben checkout out the Wasabi X SourceCred demo page you put together. Did anything else come of the main intro page and/or is that template (intro page and cred graph UI) something we can easily fork and fill with content for other projects?