It seems like there's an opportunity here for the big three.
I'm thinking something along the lines of, "Don't like the way services like SourceForge are handling your project nowadays? There are better services to use; here's a list. Obviously, we'd like you to use ours. We've already set up a home for you on our service in anticipation of your stay with us, which we think you'd enjoy. You'll find that it's already fully furnished, even. Here are the keys. Give us the go-ahead and we'll aggressively pursue the takedown of badware distributors."
The benefits to any of the three who go for this plan would be the host's association with such high-profile projects. GitHub may look at this and decide that at this point in their trajectory, there's just not enough in it for them, but it seems like either GitLab or Atlassian could benefit from it.
At GitLab we already have one-click importers for GitHub.com, Google Code and Bitbucket. We would love for someone to contribute a SourceForge importer.
That's not quite the angle he(?) was going for — he(?) is saying that an aggressive campaign from one of the Big Three in Git hosting to aggressively take down badware distributors while hosting your software would be one hell of a PR campaign.
I don't think any of the big three are hosting badware. We want people to choose GitLab, I don't want to start distributing software like Gimp without their blessing.
I think you're still not hearing what I'm getting at. The idea isn't that the big three are now peddling crapware-infested downloads, but there exist services like SourceForge and tons of download sites that are.
This is about aggressively courting existing projects that may still be on SourceForge out of nothing more than inertia. Migrating away is a process, even with importers. My original comment was about surveying the landscape for potential candidates that you'd like to see using GitLab, and then go ahead and set up a home for select projects before approaching them. This could include reserving accounts for the core developers, pre-seeding the project with whatever importing would be required, and just generally making it stupid-easy to migrate--as easy as just saying, "yeah, okay; we'll do that", and then setting up their password.
If you're worried about doing anything with their blessing, this could all happen in such a way as to not be publicly accessible until the project actually gives the go-ahead and confirms they would like to make the switch.
Making it easy is a great idea and our on-click importers are getting better all the time. Pre-creating all SourceForge accounts and content is wasteful, many good usernames will go unused and all our backups will contain many projects that are never accessed. So we'll focus on making in the import good and fast instead of doing it in advance and emailing people about it.
