Imagine you're building a super cool LEGO castle with your friends. This algorithm is like a special scoreboard that figures out who's helping the most and how important their help is.
First, if someone helping is a robot (like a "bot" that just does automatic updates), we don't give them a score. They're just doing their job, not really contributing like a person.
🛑 They get a special
-1score, meaning "not a real person."
We start by giving you points for a few things:
-
Commit Count
This is like how many times you added a new piece or changed something in the LEGO castle. More changes = more points! -
Lines Added
This is how many new LEGO bricks you put in. Building new stuff is super important, so you get good points for this. -
Lines Removed
This is like taking out old, broken, or unnecessary LEGO bricks. Cleaning up is also helpful, so you get some points for that too! -
Net Lines
Basically how many new bricks you added after taking out any old ones.
We also care about when you last helped.
-
If you helped very recently (like in the last few months), you get a bonus!
It’s like extra credit for being on the team right now. -
If you haven't helped in a while (like a year or two), your points start to "rot" away.
Like a plant wilting if you stop watering it.
The longer you're gone, the more your score goes down — but it never hits zero.
We also give you a little bonus if you've been helping with the castle for a long, long time.
It shows you're committed!
Sometimes someone just adds one LEGO brick and then disappears forever.
We call them "drive-by" helpers.
- If someone only has one "last help date" and "first help date" (meaning they only helped once), and they were supposed to get a big recent-helper bonus, we give them a small penalty.
- We don't want to "kick them while they're down," but we also want to be fair to people who stick around.
We also have a rule to make sure no one tries to cheat the system.
- If someone makes a ton of tiny changes (like adding just one LEGO brick 100 times), we check if they're actually doing real work.
- If their changes are super small and super frequent, they may get a penalty.
We want big, meaningful changes — not just lots of little ones to trick the scoreboard!
Imagine you have a few friends who are super good at building LEGO castles.
- One friend, "Jaex," built like a few million bricks
- Another, "BrycensRanch," built maybe 250,000
Their raw scores would be huge and really far apart! To make it fair:
- If your score is ≤ 1000, you keep your exact score.
- If your score is > 1000, we use a special math trick:
"Okay, you're amazing, but your score will go up slower now."
So even if Jaex built a few million bricks and BrycensRanch built 250k, their final scores won’t be insanely far apart.
This keeps the scoreboard fair and motivational for everyone!
That’s how we figure out everyone’s special "helpfulness" score for our LEGO castle! 🧱✨
So, we’ve got our helpful builders, but what about the plugin publishers?
Some people publish amazing stuff and others... well... they try.
This score helps us figure out how trustworthy a publisher is — like checking if their LEGO kit won’t fall apart when you pick it up!
We look at how active, popular, and well-reviewed a publisher's plugins are. Here's what counts:
- Plugin Uploads: Publishing more plugins means you’re trying to help. Big points!
- Plugin Downloads: If people are downloading your stuff, that’s a huge sign of trust.
- Plugin Reviews: If your plugins get lots of positive reviews, that’s awesome. You’re making the community better!
- General Publisher Reviews: Outside your plugins, what do people think of you?
- Report Outcomes: If you report bugs or issues and they turn out to be right — bonus!
If your reports are wrong or spammy? Not so much.
We don’t just count raw numbers — that wouldn’t be fair to smaller publishers.
Instead, we use logarithmic scaling (aka: "more = better, but not infinitely better").
Then we multiply that by how important each thing is:
| Factor | Weight | Why it matters |
|---|---|---|
| Plugin Uploads | 500 |
Shows you’re actively contributing |
| Plugin Downloads | 450 |
Proof people want what you're making |
| Plugin Positive Reviews | 50 |
Great feedback = solid plugins |
| Plugin Negative Reviews | -300 |
Uh-oh... not everyone’s happy |
| Publisher Positive Reviews | 350 |
Good overall reputation |
| Publisher Negative Reviews | -300 |
Uh... maybe you're hard to work with? |
| Valid Bug Reports | 350 |
Thanks for helping us fix things! |
| Invalid/Rejected Reports | -450 |
Stop crying wolf 🐺🚫 |
If one publisher has 1,000,000 downloads and another has 25,000, the score difference would be ridiculous.
So we cap the impact using logarithmic math, like this:
A huge number gets flattened, but still contributes a lot.
That way:
- Newer or smaller publishers still have a shot at competing.
- The leaderboard doesn’t just become “the same three people forever.”
We combine everything like this:
Trust Score =
+ Plugin Count × 500
+ Plugin Downloads × 450
+ Plugin Positive Reviews × 50
- Plugin Negative Reviews × 300
+ Publisher Positive Reviews × 350
- Publisher Negative Reviews × 300
+ Valid Reports × 350
- Invalid Reports × 450🚨 If the final score is negative? We round it up to zero. Everyone starts with a clean(ish) slate.
This Platform Trust Score tells us which publishers are:
-
Consistent contributors
-
Well-liked by the community
-
Responsible participants
-
Not spamming or gaming the system
Basically — the kind of LEGO engineers you'd trust with the instruction manual. 🧱📘