



Linux Mint Diverting Banshee Revenue 178
LinuxScribe writes "According Linux Mint founder Clement Lefebvre, the popular Linux Mint distribution has changed the Amazon.com affiliate code for the Banshee music player so that Mint, not Canonical or the GNOME Foundation, will receive the revenue from MP3 sales through Banshee. Though a trivial amount of money ($3.41 in November 2011), Linux Mint's actions still raise the question: how should revenue be shared between upstream and downstream FLOSS projects?"
Re:Nothing really (Score:2, Informative)
According to the article, the link is not dead (even though Mint thought it was). It just works only from within Banshee.
Re:Find a better case for the discussion (Score:2, Informative)
Per the GPL what they are required to do once they make their change to the affiliate link is make the source code available.
We're not talking about what they are required to do, but what they should do.
GPL covers copyright law, not ethics and the human heart.
But we're not talking about the GPL, but about ethics.
I'm just shocked (Score:4, Informative)
Re:They deserve it (Score:5, Informative)
Not in the project's scope - Wayland is a display server; it makes sure that crap in memory buffers wind up on the screen in the correct order. You can always run an X server to get your network transparency (this is, after all, how Wayland is first going to be used).
Sigh. No. Wayland's stated goal is to replace X. See for instance talks by the main authors about "Life after X". People are now making toolkits that target Wayland directly. History shows that if network transparency is not built into a windowing syatem from the beginning then it will suck. Dodging the issue by claiming it is not within the scope will not make the final result suck any less.
Not necessaroly
How did you work that out?
Hung applications have immovable windows which get in the way and make life suck, like other popular operating systems
Use your brain. If applications are responsible for decorating windows and therefore generating their own move/iconify requests, then a hung application won't respond. Just like hoe huing applications have immovable windows on OSX and Windows.
Impossible to use a decoration free tiling window manager to maximize screenspace
Impossible to use a window manager which adds useful extra window decorations and functions
[citation needed]
Again use your brain. If the application draws the decorations, how will the window manager augment them or completely change them without resorting to awful bodgery?
Clipboard functionality is not trivial to implement in a robust and interoperable fashion. Copying and pasting is, for the time being, still hacked and duct-taped together under all major Unix desktops.
[citation needed]
Actually, you're talking rubbish. I have implemented copy/paste in xlib which interoperates with every program and datatype I tested it with. The X11 copy/paste mechanism is actually really sensible and well designed. It goes something like this:
Prog 1: I have the clipboard!
Prog 2: I want to paste.
Prog 1: Well, I can offer you this list of datatypes which are now mostly MIME types
Prog 2: Excellent. I'll have image/jpeg, please.
Prog 1: OK, then. Here you go.
That's basically it. There are some minor wrinkles, like the "here you go" part having a mechanism for chunking the data so the server doesn't have to hold it all, and programs don't have to do special things to avoid hanging with large pastes over slow networks. But basically, it's simple, robust and effective. It's also sufficiently flexible that XDnD was added without any server or API changes, just using existing mechanisms. And it's also sufficiently flexible to allow the sort of persistent clipboards which exist on other operating systems, which work as follows:
Prog 1: I have the clipboard. ... same negotiation as above ...
Clipboard manager: Gimme everything.
Prog 1: OK.
Clipboard manager: I have the clipboard! I can offer ALL THESE datatypes...
Again, simple, ffective and robust.
It's going to bring internal overhead to a minimum, by letting the kernel manage the hardware, take care of double-buffering and minimize the amount of work needed to actually draw anything - have you ever actually tried to write an X application?
Yes. I've spent more time messing around with Xlib than with toolkits.
You could also try taking a look at smspillaz's blog, where he regularly pulls his hair out over some brain-dead functionality or unexpected race-condition and deadlocks caused by X..
That's curious. Given that X11 is single threaded, I wonder how he gets race conditions. I've never found one.
Oh, and it's still not possible to get vsync working on a multi-monitor setup (and not even all single-monitor setups) under X.
Works for me. The problem with more than one monitor is when they run at different refresh rates. There's nothing inherent in X which makes it any le