The article was published in 2010, and GitLab—an open-source distributed version control system based on Git—was launched in 2014 [1].
I would like to know the author's thoughts on GitLab now.
Do you think GitLab is underrated?
back
2 comments
Gitlab is slow and bloated but otherwise better than GitHub in almost all things, especially the CI.
GitLab is open core not open source. Forgejo is true OSS.
Is that actually a useful distinction? All open core means is that a project has a true OSS core version plus non-open parts that provide extra features beyond what the OSS core offers.
It can be the case that the OSS part of an open core project has the same or more features than a competing project that is all OSS.
There is an inevitable conflict of interest when upstream is open core.
Some features are proprietary, so users of the free edition can't use those features. But upstream also doesn't want to _improve_ the free version by including those features.
For example, mirroring specific branches or a remote repository requires a Premium licence. Filtering branches based on a regex isn't a crazy complex feature, so somebody interesting in using it could potentially implement it and send a patch. But upstream has no motivation to accept such a patch: they already have this feature in their paid version.
The distinction between open source and open core is (somewhat generalising) that an open source project will take improvements and fixes, but an open core project has motivations to reject specific improvements, because they charge a premium for those improvements in their proprietary offering.
The constant spam from Forgejo supporters who "are the only true OSS" is very tiring.
I don't understand the distinction. If I can get one car for free that doesn't have a car seat, or I can get another car for free for which I can get a car seat (at a cost)... why is the first one superior? Why is the "no paid addon" a good thing if it doesn't have the addon at all? It obviously doesn't benefit people who need car seat, but does it even benefits the ones who don't?
The core of "open core" products is open source, as per the OSI definition. What do you all want?
It has to do with the way power and incentives are configured within the project, and therefore what can be expected of the maintainers in the future.
For some people/use cases, the threat of developers rug-pulling a tool you depend on is not a big deal as long as it's good right now. But in many situations the tool which has less features but also less incentive to rug-pull wins out.
Anyone can "rug-pull" a project, whether it currently has non-free features or not. You can't retract already-published versions, but anyone can make non-free plugins or forks for existing MIT-licensed code (GitLab and Gitea are MIT).
I guess some might think that because they do non-free parts now they are likely to make more of it non-free later, is that the argument? If yes I don't really like this Minority Report approach to judging projects for what you think they might do.
> because they do non-free parts now they are likely to make more of it non-free later, is that the argument
Yes, that's one indicator of how the incentives are structured, though there are other factors to consider too - mostly regarding where the money comes from and who is involved in the decision-making.
Perhaps you find it dystopian that people make predictions about future behavior and use them to inform their decisions about who to trust. It's very common though, and is the basis for the concept of reputation.
The parent post highlights one of the core reasons behind the failure of FOSS: its growth was built on a base that cared only about the gratis aspect without any real interest in the libre aspect. They're perfectly happy to take any FOSS they can get for free and pay for any proprietary software needed with the money they saved.
If you have to do something in exchange then it's not gratis OR libre. Whether that's enforced by law or community shaming.
Are there any key features which is not in the open source version?