▲ 186 points
back
102 comments
I wonder if this goes against github's "no automated starring" policy [0] (that people have been banned for in the past). Maybe "coordinated inauthentic activity" ?
Either way, feels sleazy.
[0] https://docs.github.com/en/site-policy/acceptable-use-polici...
I fail to see a reason where it would be a violation because of the following reasons:
1) it's a response to a user's request, i.e., not initiated by the repo author
2) it depends on consensus of the user.
3) it's not automated (most of the items from the policy are related to automation).
4) the repo author has no obligation whatsoever of maintaining the project. He is not paid or forced to do it.
5) if the user really wants to apply this change or disagrees with this practice, he/she can always fork it.
That said, I understand that it may still not feel "fair" compared to other projects that don't follow this practice. Or the feeling of "wanting to help but you're asked to do some things first".
Companies already do that to accept your pull requests though [0], which takes way longer than giving a star - and I didn't see a complaint about it on HN
It's quid pro quo, isn't it. The repo is buying stars with bug reporting as currency.
Which may well be against GitHub rules but more than that, it's a stupid thing to do. Bug reporting helps projects get better. They're blocking real bug reports until they get a star. Seems like self-harm to me.
Regardless of the current rule, it really ought to be a violation. If it's not, all repos that care about star rankings are forced to start doing the same thing to compete, which eventually turns stars into a metric of which repos have the highest quantity of annoying problems that users want to report, rather than a metric of whatever it's supposed to measure.
It is at least partially automated. The issue was closed by a bot which requires a star to re-open.
This cheats the other users (like me or you) -- if we are looking at the project's star count, you are likely trying to judge project's popularity and get a measure of how many people like it.
And dae's star count is basically a lie - it does not represent how many people liked it, but rather how many people had found annoying bugs in it. Someone who is forced to use dae and hates it would still need to star it. And as you said, other projects who don't have this practice are in disadvantage.
(And that's why I have no problem with things like CLAs, or requiring details bug reports: they are not user-visible and they don't mess with global rating.)
There's a submission regarding CLAs nearly every month. But complains regarding them is not for the time it takes to sign one. Similarly complains about the practice here aren't for the time it takes to star a repo.
Gaining a permanent reputation for petty authoritarianism, even with contributors who clearly want to help
versus
gaining one (1) point on GitHub
Truly craven that this is coming from an automated bot. The extent to which social media likes have broken peoples' brains is something to behold.
Craven is definitely one of the words I'm groping for to express just what is wrong with this.
I really don't understand the "seems fair" comments.
I (stupidly I see now) treat the stars as feature just for me, like a bookmark in my own browser. There for my use, not anyone else's.
Obviously that was silly since they are public and anything that is public is someone somewhere's currency, and anything that is currency will expose the lowest of the low human behavior.
And this IS automated. The automation is in the form of a policy rather than a scripting language, but it's still a mechanistic rule where a star is produced by application of an if-this-then-that, and by a transaction, purchased exactly like with money. By at least those TWO different means this is not an honest reflection of a users regard or admiration or expression of value.
Github surely would see that as devaluing stars and harming a feature of their site. Surely github does not want stars to become like amazon reviews.
Are you arguing that people should not be authoritarian on their own personal projects? To me this signals having healthy relationship with the project and knowing how to set boundaries (small as they may be) to avoid maintainer burnout.
My point is that if you're an organization looking to use a tool like dae, the maintainers' behavior here is a huge red flag. Of course open-source projects can do basically whatever they want. But dae is really not a "personal project," its clearly designed for use by the broader community and the maintainer clearly wants lots of people to use it. If I am considering using any open-source software, I would want to understand the devs' decision-making behind denying good feature requests:
1) The best possible reason is work-life balance, "hey I'm not getting paid for this," etc etc. I am not demanding that dae's maintainers accept the change simply because it's a good idea.
2) An unpleasant but probably acceptable reason is obstinance, stubbornness, etc. Maybe you're wrong, the feature is a bad idea. But even if you're right, sometimes you have to accept the aesthetic/ideological quirks of the devs.
3) The worst possible reason for denying a feature request is pettiness or narcissism, which is exactly what dae is doing. Perhaps the user had legitimate reasons for not starring the repo (GitHub "learns from" your stars and will suggest related repos in discovery, which can be annoying). But the idea of labelling CI tests as a "wontfix" until you get a stupid GitHub star is just horrendous open-source management.
understood
I don’t think I have ever starred anything on Github. I think in this sort of situation I would add the star and remove it once the fix is added.
I'll star a repo when I want to add it to my "favourite" or "interesting" repos - they are all listed under "Your stars". Is there a better way of achieving the same?
I'm using a local gitweb instance for this purpose. I have a directory of favorite repos cloned using "git clone --mirror" and a cron job to sync them every 3 hours. gitweb provides a web interface for browsing them.
I like that it lets me open files without JavaScript enabled and search the code without logging in. I still use GitHub.com to read the issues on a couple repos though.
Example of what gitweb looks like: https://sourceware.org/git/ - There's also cgit which is similar: https://git.kernel.org/
is lack of privacy your issue with the "your stars" page? consider a bookmarks manager.
In general those things seem like site-specific bookmarks. As a general principle I want to just use one bookmark feature for all web things.
That just seems like petty behaviour.
Why?
I'm of two minds on this:
1) Good: This adds "just" a touch of friction to things. You can't just do a drive-by on the project like so many thanks to ill-advised "social good" corporate review policies (screw you, Google, for pioneering that). You have to at least star it which at least signifies some level of commitment, identity, and existence.
2) Bad: This causes an inflation war between those projects that do this organically and those that force people to star them. It squashes the signal in the idea of stars.
Github is going to have to remove identity from the stars if they want to stop this from descending into chaos.
You think stars before had some widely agreed upon common idea that was not chaotic?
I agree with this. If you use some software (and its author cares about internet points), you can at least give it that star. It's better than paying with money, and if you are requesting a feature it makes even more sense to pay with a single click of a mouse. Show some appreciation, people
It seems ludicrous to have to star a project for lacking a feature you need. What's next, starring projects for having glaring security problems?
No. I'm not going to break my own user pattern (I use stars as bookmarks, that's it. and I unstar pretty aggressively once I've done my business) to satisfy someone's ego. I'd much rather give a donation or bounty than go through that emotional manipulation. At least giving money is a business-neutral way to show demand for a feature (for a product I may or may not like).
Also, as a rebellious streak I am much less likely to star something that already has (proportionately) a bunch of stars.
Seems obnoxious without a doubt, and particularly egregious if reporter is bringing/offering a fix. But maybe I should try to be open-minded about this? Issues that provide even badly written bug reports are probably doing a service to the project itself. Feature requests OTOH might just be internet strangers requesting free labor on someone's generously-public yet resource-starved passion project. In the 2nd case, and without speculating on why the project owner would even care.. maybe fake internet points in lieu of other payment is the least people can do. That said though, obviously if everyone does this, it will erode the signal of what stars are supposed to mean. Hmm.
> if everyone does this, it will erode the signal of what stars are supposed to mean.
What signal? To me, starring a Github project means "I'm interested enough in this project to want to be able to find it again". That's a significantly lower bar than the poster had here, of "I've a new feature proposed for this project, and I'm even interested in implementing it myself if the maintainer agrees".
There's a loose correlation of stars with quality/popularity/ activity. All of those correlations should become stronger by having active users and bug reporters star the project.
Not sure which is more gross: having a bot farm for imaginary internet points, or the people defending the practice.
No clout, no work seems reasonable if on the nose. At least you don’t have to subscribe to a newsletter.
Imagine if I could demand that only people who have upvoted my comment could reply to it? I would get much less disagreeable replies.
This is unethical if (1) the condition was not made clear before the issue is written, and (2) the user cannot delete their report before it is published. (It is not in the current issue template, and I don't know if it was there.)
1. The issue writer will invest time and efforts, and publish their findings. Nobody's time is free: the maintainer's time is precious, and so is a user's.
2. Now, _after_ the issue is published, the maintainer additionally asks for a certain condition (give money or a star) to be satisfied.
What if the issue writer does not want to give that thing? The maintainer is now in an unfair position: they can still read the published issue, improve their software, but is not obliged to give any feedback or even credit the reporter.
It would be fairer if the condition was clearly conveyed to the reporter before they write any words. The system should simply not allow these issues written by the dissents to be created in the first place, and in that case the funny duck would also not appear before our eyes.
What are stars even worth? Certainly not the number of possible contributors this chases away (I see issue reporters as contributors as well since they’re spending effort finding/debugging what you would have to do at some point anyways)
All this for a measly 1500 stars
Maybe to put the repository on a resume. Github also gives out free Copilot licenses to active contributors of opens-source/free-software projects but they don't say specifically what the requirements are. Maybe if you own or contribute to a repository with many stars you're eligible for that.
It helps make the repo most popular
- more bug reports
- more community documentation
- more user testing
- etc
Reminds me of how how people give trivial stuff away for "free" in online marketplaces, but ask to be paid with a good review before handing it over. Do that a few hundred times and whoa, you're an amazingly rated seller!
Yeah, that's why I hate this model to begin with (and basically never leave reviews for software). If you have to ask/bribe me for a review, you ruined the point of what a review it. It is to let other prospector buyers know my genuine experience with the product or seller.
I don't mind being prompted for a review after the fact (because sure, I will forget otherwise), but it should be a prompt, not a beg.
This is Trump demanding that everyone at a meeting say something good about him.
A star is me saying "I admire this".
This is saying "hey you, admire me."
Super gross. How is anyone here trying to say that's reasonable?
And aside from being low, forget that, it's also stupid because it hurts your own self. Some users do submit feature requests in a demanding entitled manner, but most don't and even those that do are all still useful information about what the users want.
This user is only submitting an idea for consideration, and offering to do the work themselves if the idea is agreed with.
The devs* response to this is utterly bizarre in that context. "I won't hear your idea unless you like me." ??? The user is trying to GIVE them something!
Not all gifts are desirable of course but the value or fitness of the pr was never even considered in this case. They are refusing to even look in the bag to see if it's a turd or a gold brick unless the user claims to like them.
* devs bot, aka, which would seem to pretty squarely hit the acceptable use terms that specifically covers stars and automation and authentic user interactions. Not only is the devs policy a form of automation in that it's a plain mechanistic rule no different from a line of code, and is also a transaction like buying amazon reviews, on top of that it's literally a bot implementing the policy!
The mean goose avatar for the bot is perfect.
That's stolen from "Untitled Goose Game".
This is ridiculous. Not every issue means someone likes or even uses the codebase.
Automated scanners, fuzzers, security researchers, and people who found an issue via a dependent codebase, are all examples where an issue might be filed but the person otherwise has no interest.
What a nonsense thing to have. It's not harmless, especially for large projects. It'll discourage and even runs the risk of artificially covering up potential security issues.
You are assuming such drive-by issues are actually valuable; to me it is not 100% self-evident, especially when most projects do not have shortage of issues anyways.
Don't forget, nobody prevents the user from opening an issue. If it's critical and the maintainers care, they'll work on it with or without the user's star.
The reactionary way to respond to this seems to be to think this is somehow authoritarian, internet point gathering or just petty.
What some might not realize is the sheer volume of requests that maintainers of popular open source software projects have competing for their attention on a daily basis (I do not speak from experience but from reading/listening to reports from those that do).
Adding a bit of friction to any of those attention grabbing things, like here through the use of an bot, is a good way of reducing that noise at least somewhat. It isn't a perfect system, but it is the one this maintainer choose to use.
>understood
iconic
Why the negative sentiment against this? It doesn't cost people anything to star a project. Whether it is a PR or not, the maintainer still had to review and maintain the code and that is time they will spend. The more stars a project gets, the more support it gets. It's not like you pay for stars or it is burdensome to click on the star button.
Because it artificially bloats the star count so that it no longer represents how many people are interested in the project.... Although if you've filed an issue you kind of are invested already
I have a family member who would get upset when others would not like her Facebook posts and would message us about it. (It was one of the reasons why I stopped using FB.) OP's submission reminds of this family member.
Because people don't like to drink a verification cans.
> It's not like you pay for stars or it is burdensome to click on the star button.
Imagine being allowed to comment here only after [starring] something. It's not like you pay for [stars] or it is burdensome to click on the [star] button.
How do toy magically get more support if all projects simply demand more stars and get them?