[hodgesrm]

> How could you control code quality if you let randos write your stuff? You can't.

Huh? You use the pull request mechanism on Github. That's what it's there for. If people want to contribute, the PR needs to be up to standards. Just to pick one example, Elastic uses this method for ElasticSearch. [1] ES gets PRs from hundreds of unique contributors each year.

[1] https://github.com/elastic/elasticsearch/pulls

[mimixco]

Pull requests don't mean the code gets integrated. It's just someone sharing a suggestion. Unless that suggestion happens to be on the project's to-do list or it fixes an open bug, I'd say it's unlikely to get merged.

Elasticsearch's top contributors (the three I checked) seem to work for Elastic, which goes to what I'm saying. No profit-making business can have its workflow dictated by the public. Even if a pull request is accepted, the team has to go through all the same code review and testing processes as if the code had originated in-house, and that's not always feasible or on the timeline when the request comes in.

[hodgesrm]

> Unless that suggestion happens to be on the project's to-do list or it fixes an open bug, I'd say it's unlikely to get merged.

There are lots of open source projects that contradict your assertion. Here are 3 projects that have been quite liberal about accepting PRs from outside.

* MySQL - MySQL AB was very open about accepting external patches. Oracle still does today.

* ClickHouse - Yandex team has the most commits, but accepts PRs from hundreds of outside contributors.

* Superset - Apache project that is supported by Preset. Again, pretty open to new PRs as far as I can tell.

Overall it's a question of how the community is governed. A lot of us view open source communities as sources of innovation. That being the case, why would you refuse a good PR? Somebody just solved a problem for you, maybe one you didn't even know existed.