Comment by domenicd | Hacker News Reader

Comment by domenicd | Hacker News Reader

Parent Full thread

I'm cautiously optimistic about this.

This is what Hacktoberfest should have been from the start. A company incentivizing pull requests against your repo should require your consent. A maintainer needs to affirmatively sign up for this sort of work; they can't have it thrust upon them. As a maintainer, all the other half-measures people have suggested (e.g., accepted PRs only; account age measurements; filtering out README changes) do not actually reduce the unasked-for burden on me.

That said, I'm not sure this year can be salvaged. It's not clear whether all the existing spammers will get the message. I hope DigitalOcean emails all folks signed up to notify them of this change. (They haven't communicated any of their changes so far in such a way.) Notably, a day ago, the DigitalOcean staff thought this wouldn't make a difference. [1] And it's even possible this will just result in issue spam begging maintainers to add the Hacktoberfest topic.

But it's also possible that this will work out. The best version of this is that low-quality contributors looking for a t-shirt get directed toward honeypot repositories, and folks interested in making substantial contributions will continue using issue labels like GitHub's own "good first issue", or the Hacktoberfest repository topic, to find places to contribute.

We'll see in a day or two whether this is enough.

[1]: https://news.ycombinator.com/item?id=24648285

> That said, I'm not sure this year can be salvaged. It's not clear whether all the existing spammers will get the message.

I think the biggest problem DO will have to address somehow is that Hacktoberfest now has become such a big event that they are getting attendees which they strictly speaking do not want to have, and the liberal “everything is open for everyone”-approach is now too open for the “game” to work.

Finding a way to encourage (new) people to contribute to open-source projects while at the same time keeping the bad people out (opportunists, spammers, etc) is not going to be easy. Locking things down (like only opt-in repos count) will severely limit genuine contributions way beyond just spam.

Whatever they end up doing, I suspect we will see Hacktoberfest 2019 as peak Hacktoberfest and from now on the event will be a mere shadow of what it used to be.

> Finding a way to encourage (new) people to contribute to open-source projects while at the same time keeping the bad people out (opportunists, spammers, etc)

This is the challenge of our time, TBH. From the Eternal September onwards we’ve struggled to sustain community quality in the face of infinite connectivity. You start open & welcoming, but eventually you get overwhelmed with new folks and you can’t socialize them fast enough to preserve what was attractive about your community in the first place.

This is the work of institution building. It’s not easy. The best examples I’ve seen are HN, Lobste.rs and some focused subreddits. All of these have strong mod teams and spam controls.

> A company incentivizing pull requests against your repo should require your consent.

Why?

Essentially Hacktoberfest turns repo maintainers in to human spam filters who decide if Digital Ocean should send someone some marketing swag. Companies shouldn't be doing that to every open source maintainer. It isn't fair on them. That should only happen if the maintainer is happy to do work on Digital Ocean's behalf.

Do you really have to ask "why"?

I do. It sounds really weird to me. The whole concept of pull requests is that anyone can make them.

Yes. Anybody should be able to make pull requests.

But incentivizing people to make pull requests is a different matter.

Should companies ask for permissions before they employ someone to send pull requests to public repos?

I think the problem was not as much incentivizing pull requests but incentivizing low effort pull requests and outright spam.

How so? Incentivising people to help (via PR) is great. It's like incentivising people to donate blood or similar "everyone benefits" scenarios.

(Yes, of course, blatant spam doesn't help anyone and is disgusting, I'm talking about in general)

The problem is people aren't incentivized to help, they're incentivized to get free stuff. So people who have no interest in helping are doing the bare minimum to get free stuff, which is a much larger number than the people actually wanting to help.

In your scenario, it would be like incentvizing "donating a vial" and hoping it will be filled with blood, but 99% of people are donating an empty vial.

People who donate blood are pre-screened to ensure their contribution won't be harmful.

Are you unfamiliar with the problems this has caused?

Have you been following the events of this year's Hacktoberfest?

"Anyone" does not scale infinitely: it only specifies who might participate, not how many can. Each project will collapse — lose work for one day or more — under a certain sheer quantity burden of pull requests. When people say "anyone can", they don't always mean "everyone do so all at the same time", as happened here. The term "anyone" makes no statement about the frequency of pull requests, and everyone is generally upset about the frequency of pull requests associated with a third party action (Hacktober). Rightfully so: their time is being wasted and their expectations of reasonable behavior are being violated.

I find it hard to consider a flood of pull requests ‘wasting my time’. Don’t look at them. It’s not as if you’re obliged to do anything with them.

Others have a different experience than yours, that is both negative and not improved by the “ignore them” process that works for you. I encourage overcoming the difficulty you describe.

I would argue that having your repository on github is a strong way to say "feel free to contribute". Not saying spam is OK, just that having to ask repo maintainers their authorisation seems superfluous.

Feel free to "contribute and get nothing" is completely different then "contribute get tshirt".

The people who participate are different and pull requests are different.

How are they different as long as they are not spam? Good contributions are good contributions and help, regardless if there's a crappy t-shirt or not in exchange for them

It's the incentive structure. It's effectively Goodhart's Law in action: When a measure becomes a target, it ceases to be a good measure. In this case, the measure is "number of pull requests", and the result of making it a measure is that people created an insane number of PRs that simply adjusted white space in a README or slightly reworded phrases but kept the exact same meaning. Because they wanted the T shirt, because they felt that it validated them as a developer (ignoring that good contributions would have been a much better validation).

You're clearly not aware of what's been happening the past few days. It's not good contributions or help.

They don’t, but they should if they intend to use me as part of their marketing campaign that could lead to spam on my repos and pissing me off isn’t one of their goals.

Reply on news.ycombinator.com