As someone who visited Sourforge to download GIMP 7 years ago, why does this site still exist?
2. It gave you all the stuff you needed for your project at one place. The SCM of your choice, mailing lists, hosted forums, wiki, bugtracker, and a website with PHP/MySQL. They have retired hosted forums, wiki and bugtracker now.
3. It increases the discoverability of your project. The project summary page actually gives a nice overview of the main points of your project such as licence, category and features; They have things like project of the month, etc.. Without sourceforge (and freecode) we are at SEO, social media and word of mouth for open source PR. This is much more effort.
4. They allow huge binary downloads which can be uploaded with e.g. rsync and scp.
Loosing all this might be fine for a large project such as GIMP, as they have the manpower and donation money to do it themselves. For smaller projects this distracts from developing the project.
I don't trust SourceForge anymore like I used to in the past. Stuff like taking over old projects and adding adware to them definitely don't help, but even ignoring that SourceForge still feels dead.
At this point I have to wonder - why not have a paid download service for OSS builds - like 1-2$ or whatever symbolic price minimum - this will probably cover the download costs and maybe leave your project some tiny bit of cash.
At the same time provide a free build link on torrents so that your users can access it trough torrents if they don't want to pay - it doesn't cost you anything (even the bandwidth if your project is hosted on something like GH pages).
So you pay 1-2$ for convenience or go download trough torrents or if you can't use torrents and you're brave search google for mirrors.
Source: I manage an open source project used by a few million people.
I assume it's much lower than 2%, which would fit my experience. From about a thousand users (a conservative estimate), I get around 10m$/user on average, and 0.3% of paying users.
And that is why everyone turns to the ASF for help in this situation, since they already exist, but that's probably not sustainable.
The answer is not clear.
For example if github provided this service instead of binary releases - you would receive say 50% of the income they keep 50% for bandwidth/service. If the person uses alternative means than you're not spending resources anyway.
The take away for me is that Sourceforge was a distribution as well as a management platform for open source projects. The distribution part differentiating it from other project hosting sites.
As a result of the community backlash against SF's download wrapper, I've had to click through more "attack site warning" pages in the past month just to reach some valuable SourceForge project (of which there are many thousands) than possibly all the previous year combined, though I notice in the past few weeks, Google have silently dropped SF from their blacklist (no PR fanfare this time however).
People continually use GitHub as some bastion of how things should be done, but someday GH will also exhaust its natural growth and start more aggressive upselling (regardless of their existing business model), quite possibly under a different management after the founders have all left for greener pastures. It's not a question of if that will happen, but when. Maybe in 16 years time? Who knows. It seemed deeply unlikely at one stage to imagine Google would flush Code down the toilet, but if it can happen at the second most valuable company in the tech industry, it can easily happen to a smaller company like Atlassian or GH.
I'm curious as to why you put this in the past tense. SourceForge never backed down from their actions and continue to use the wrapper.
> someday GH will also exhaust its natural growth and start more aggressive upselling
Acknowledging that GH can go the same way as SourceForge is a very healthy mindset, however, "will" is a strong word because GH has a monetization plan (is always has) - the lack of which is what ultimately caused SF to go down the path that they did.
What GH will or will not do does not deter from what SF is currently doing. Stop looking in your crystal ball and start looking at what is happening in the present.
> It seemed deeply unlikely at one stage to imagine Google would flush Code down the toilet
If you know Google you know this isn't the case. Google are ruthless with cutting ideas and products - just look at Reader. Google realized that Code wasn't making any money, saw that the platform was stagnating and saw that GH was doing a much better job at something that they had never really thought out. Exactly the same situation as SF. Two drastically different approaches to product death.
> If you know Google you know this isn't the case. Google are ruthless with cutting ideas and products - just look at Reader.
Well, yes, it's obvious in hindsight... but Google started Code in 2005 and project hosting in 2006, while Reader was shut down in 2013.
Github wasn't even launched until 2008.
Google has been like this from very early on. I won't be surprised if I wake up to the news that gmail is getting the axe should google decide that the product is no longer commercially viable.
> I've had to click through more "attack site warning" pages in the past month...
You had to click through those "attack site warning" warnings because Sourceforge was serving malware(!!), not because of community backlash.
> Google have silently dropped SF from their blacklist...
I haven't checked, but that would probably be because SF.net decided to stop serving malware.
> ...someday GH will also exhaust its natural growth and start more aggressive upselling...
And when that day comes, folks will no longer recommend Github for project hosting.
Dice's holding company completely failed to understand the nature of the community that Sourceforge served, and tried to increase revenue by attacking their customer's computers, rather than by hiring folks to rework and redesign the site, add new services, and remove any rough edges on the existing ones.
I work for Atlassian and I agree with the other posters that you're right to be skeptical, but it's worth pointing out that Atlassian has many product offerings other than Bitbucket.
Despite best efforts, there is no known corporate structure that can protect its' customers from the lack of perspective of future leaders, who will be focused on keeping as much staff employed as possible and delivering profits to investors.
The answer to this used to be places like cdrom.com, but I just checked and they're gone. :)