Linux downloads have declined because most developers get binaries from their distribution and fewer and fewer old projects (i.e. those from source forge) need to be downloaded as source.
OS X is a backwater as far as binary distribution is concerned and so developers have to use brew or macports. Every time you do "brew install blah" you have a 1/3 chance of downloading a project from source forge.
Another factor is Github. It could be that a lot of the new software for Linux is now hosted on Github, Bitbucket, and other DVCS-friendly sites. Meanwhile, as you say, the more stable stuff on Sourceforge is likely to be handled by distro-specific repositories.
I'd like to see the statistics comparing github, gitorious, bitbucket, google code, heck, even codeplex. SourceForge is no longer the only goto for compiling from source, either.
I can’t remember the last time I found a project hosted on SourceForge. Actually, I can, it was http://webcam-osx.sourceforge.net/ about 4 years ago. I assumed everyone had moved en masse to Github?
I think they need to work on their data collection. The unknown "Other" is almost as large as the other OS categories combined. This makes it hard to draw any conclusions.
They seem to have counted downloads, rather than site visitors.
The decline in Sourceforge downloads by Linux users is likely to be down to the fact that most will simply grab the package from their distro's repository, rather than getting it from SF. If I download, say, FileZilla on my Windows machine, I get it via Sourceforge. On my Ubuntu box, it's in the repo.
Yeah that's how I get almost all the software for my GNU/Linux boxen, all from repositories. And if I'm hacking on something just pull it out from Git or BZR or SVN or similar. Many projects also left SF since they started with changes that made it a bit of pain to use.
God, yes. It frustrated me so much that I ended up just copying the data into JS arrays and generating my own charts using the Google Visualization API.
Indeed. A single chart with time on the horizontal axis, downloads on the vertical axis, and one line per operating system would have been far more comprehensible.
Instead we get time inverted, bar graphs, and no way to visually compare the operating systems.
It took me a while to figure out the dates were MM/DD/YY. Even then, why is the day included? Are these per-day, and if so, why are they making sweeping generalisations about all of April? If they're just point markers why aren't they in line, and why are the bars separated?
Please note how the "other" category consists of nearly half of the downloads. I would say that this adds a lot of uncertainity in this dataset. Without knowing the distribution of the OS-es that shows up in "other" this data is pretty much meaningless.
Good to see data ming on this... it's been obvious in my work circles through 2011 so far. "Real stuff" is getting done on Mac OS, not just visual and content.
That was not a good article at all. The title is misleading and the content has little to nothing to do with "OS Wars". We all know the users are benefitting from the 3 big competing OSes. Then they go on to basically say "hey, look at us, we're still relevant! We have some Analytics data that you should read as coming from a major authority".
I have nothing against SF at all but this kind of sounded like a call for attention. They talk about this data like its the final word on the subject but leave out so many other really popular resources where people get their apps like App Stores, Package Managers, GitHub, etc.
And what OS wars? This is 2011 not 1997. The OS makers are competing but they're not trying to be all things to all people like in the old days. We've got Linux distros for all types of users from sys admins (Arch, CentOS) to casual users (Ubuntu and other Debian based distros). The Mac has its own uniqueness that appeals to a broad audience and some people want their old standard, Windows. Some are gaining ground, others losing, but it's not like 10 or 15 years ago where people were looking for a winner. Some people are still and I'd say that's kind of dumb and jingoism has a bit to do with that.
This article should have had the title "SourceForge Sees Major Increase in Downloads from NameofOS". I call link bait.
Actually, I expect that Mac users are overrepresented: every time you go "brew install blah" there's a good chance you're downloading off of SourceForge.
Surprisingly,
since 11/1/10, windows has less downloads than all others combined.
Since 3/1/11, windows + mac + linux have less downloads than other alone.
The difference between the OS distribution on SourceForge and the OS distribution in the real world only shows how irrelevant SourceForge has made itself.
OS X is a backwater as far as binary distribution is concerned and so developers have to use brew or macports. Every time you do "brew install blah" you have a 1/3 chance of downloading a project from source forge.