-
Notifications
You must be signed in to change notification settings - Fork 5
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Reviewer Incentives #60
Reviewer Incentives #60
Comments
Source Specificationuusd.ubq.fi#3Taking this large pull as an example: https://github.com/ubiquity/uusd.ubq.fi/pull/3/files
We can exclude the lockfile diff and other common generated files line changes count using linguist-generatedI think a minimum credit of $25 could be nice because even taking the time to see a tightly scoped pull of only a few lines requires context switching and thinking about the changes in relation to the codebase, also while checking CI status and conversation state. We could add this minimum ONE TIME if the reviewer left a conclusive (approval or request changes, not just commented) so the full review would yield Handling Multiple ReviewsProblem: pulls may be reviewed multiple times by a team member. How do we handle this dynamically (no hard numbers) so that the incentives scale with small and large pulls accurately? Solution: we track exactly which lines the reviewer is reviewing and credit only for changed lines reviewed. Example:
Second example, starting from the middle of the commits:
Footnotes
|
A last consideration is the premium for priority level. Similar to our other rewards, we should grant higher rewards for higher priority projects. Multiplying the entire output doesn't seem right, because $25 * 4 (priority 4: urgent) = $100 just for leaving a rejection or approval which seems excessive. In example 2 above, as this was a priority 4 task: $92.46 * 4 = $369.84 + $25 = $394.84 in review rewards for me. Basically the same for rndquu as they approved the last commit at the time of writing. whilefoo last reviewed ubiquity/uusd.ubq.fi@7e705c1 so
zugdev ubiquity/uusd.ubq.fi@f07a96b
This does not include the conversation rewards so we might need to normalize them to be the same amount for every contributor. And review rewards only for collaborators. |
/start |
/help |
Available Commands
|
/start |
Tip
|
Could you please create a repository that I could target with this? |
I agree this would be great and solve a problem that's existed for some time now, particularly the last couple months although pace seems to be improving rapidly these day in my experience. I started on https://github.com/ubiquity-os/ubiquity-os-plugin-installer ~3 weeks ago and it will be tied away in the next week so I am free to pick this up and rush it through if needed, lmk ty. |
@kingsley-einstein, this task has been idle for a while. Please provide an update. |
On it |
@kingsley-einstein the reason why it follows up is because you aren't working on it. You should be working on it to avoid the follow ups. |
@kingsley-einstein, this task has been idle for a while. Please provide an update. |
I'll push in a moment |
@kingsley-einstein, this task has been idle for a while. Please provide an update. |
Currently working on this |
/help |
Available Commands
|
/help |
Available Commands
|
/query |
|
/query @surafeldev |
|
/wallet 0xB13260bfEe08DcA208F2ECc735171B21763EaaF6 |
+ Successfully registered wallet address |
/query @surafeldev |
|
/start |
Tip
|
can I change the registered wallet, It seems like it can`t be changed by re registering my wallet address again ? |
The spec basically gives a reward to the diff of the commits combined between each review if this were to be a separate plugin wouldn't the reward infrastructure have to be implemented first? |
This seems like it might be related to that one problem a long time ago with the shifted rows in the database @whilefoo @gentlementlegen any ideas? Also @ishowvel that is an accurate assessment so I think for the short term it would make sense for @gentlementlegen to implement a module inside of the text-conversation-rewards plugin and then later we can move it out to its own plugin once we support "permit requests" from all plugins to the kernel and the kernel generating them all in one shot. |
I think this pull-request should help re-registering a different wallet to an account. |
/start |
! This issue is already assigned. Please choose another unassigned task. |
Maybe this can be done after this task but I think we should also incentivise fast reviews, either by increasing the reward for fast reviews or decreasing reward for slow reviews. |
I remember thinking of this in the past but I couldn't find my old proposals. It would be helpful if you put together a proposal with the math formula as thats the most important part to figure out. |
@surafeldev, this task has been idle for a while. Please provide an update. |
Passed the disqualification threshold and no activity is detected, removing assignees: @surafeldev. |
/start |
Tip
|
I have been researching and find out how should it be done, I have started working on it. |
@0x4007 |
That looks fine for now, we can keep iterating to find something good. It's kind of a minor detail. I think the ID is not useful to show so you can exclude it. |
The current sumRewards function gets away with choosing the least of two values to limit rewards for each type of reward because theres only one type of reward by adding this pr we now have multiple types of rewards 25.006 (review diff + review base) this would become 6.1 which would go over the limit, was this behavior intended? |
Reviews always seem to be slower than new pulls being submitted. Perhaps with better incentive design we can speed up this operational problem.
My original spec was dense but provides a lot of context on the problem. I asked Claude to rewrite it to be more clear, but in case you want more context, please scroll to the bottom.
Jump to this comment to see final output amounts from a large pull.
Claude Rewrite
Base Calculation
Core Components
1. Minimum Review Credit
conclusiveReviewCredit
)2. Line Change Calculation
(additions + deletions - generated_files) / 100
3. Multiple Review Handling
Example Scenarios
Single Complete Review
Incremental Reviews
Implementation Notes
File Exclusions
.gitattributes
withlinguist-generated
yarn.lock
Diff Comparison
...
)Review Tracking
The text was updated successfully, but these errors were encountered: