Somewhat related, an indie company managed to get a fighting game out on PSN and XBLA. It just got an update on PS3, but it can't get updated for Xbox because the patch is bigger than 4MB. [1]
Keep in mind, these are free updates that fix bugs and stability issues, not paid DLC. Even so, patching is really, really expensive. [2]
It's complicated. Wayland can replace some parts of X11 in Linux distributions, but it relies on some other parts of X11 (like libxkbcommon).
I believe Wayland window managers [1] will replace most of X11 for most users in a couple of years, but X11 will still be around to run legacy apps (similar to the way you can run X11 applications in
I also think that Wayland will replace X11 in the way that most of the X11 developers will move to Wayland. X11 will still be around, and probably still be supported by window managers for quite a while though.
1. Wayland is actually the protocol/API, which window managers/compositors and the applications that run in them will implement. The reference compositor for the Wayland project is called Weston.
>I believe Wayland window managers [1] will replace most of X11 for most users in a couple of years, but X11 will still be around to run legacy apps (similar to the way you can run X11 applications in
This comment must have been submitted from an X11 application running in MacOS X. I am having the s
I believe Wayland's stated goal is exactly that: to replace X11. of course they don't mean they're going to wipe out X11 from existence, if you prefer that you'll still be able to use it, but wayland should eventually be able to completely fill the role that X11 currently fills.
Very few apps call the X api directly. Nearly everything uses GTK or Qt, which they'll still be able to do, except that behind the scenes GTK/Qt will hook into Wayland on Linux and X elsewhere.
> That doesn't sound good. It seems that Wayland will have a hard time to replace X11 in the next decade. Currently I see no reason why I should switch.
The reason you will switch is that plain X11 goes unsupported and will not work on modern systems.
The initial thing you will switch is not Wayland proper, it's just X11 sitting on top of Wayland. Thinking of Wayland as a competing display server is as of now not really realistic -- it's much better to think it as a refactoring of the internals of the system that uses X and compositors, taking the parts of both that need privileges on the system and really fast communication and merging them together in Wayland, leaving the rest of the parts into X.
Wayland is going to replace portions of X11 pretty much as fast as it can. Basically, the X11 devs want to drop all the parts of the pipeline that Wayland implements from X11 and move to (usermode) X11 on Wayland.
Yep. While it has it good points, it is never going to replace x11 on my laptop since it requires compositing. And i more or less had to turn off compositing on my thinkpad edge since it reduced the akku lifetime drastically and was very noisy.
I don't know why they don't, but I'd give serious thought to not supporting Windows at all if I had to use mingw/msys. And this is coming from somebody whose first order of business on a Windows machine is to install cygwin. The environment is just Not Friendly.
Visual Studio is the vendor-suggested way of building C++ and it's free besides; there's not really a good reason not to use it.
I've found mingw and msys to be very fast and easy to use.
In my experience it's been Cygwin that's a bloated, non-intuitive, awful monstrosity to be avoided at all costs. I've had great difficulties in the past with Cygwin DLL's, particularly when a program comes with its own Cygwin DLL which is a different version than the system Cygwin.
I haven't used Cygwin since 2005 or so due to these difficulties.
For Windows virtualization solutions, my first stop these days would be Virtualbox, qemu or the like. Second preference is mingw/msys.
I've also had a positive experience with a little-known solution called Colinux [1], essentially a Windows port of another little-known technology in the Linux mainline kernel called user-mode Linux. Colinux requires some setup, especially if you want graphics (for GUI work, you need some remote desktop with a Windows client like Xming or VNC). Again, I'd recommend VirtualBox or qemu for casual use, but Colinux is an interesting technology, and I've found it gets very good performance.
My negative experiences with Cygwin were so great, it is only something I use when there is no alternative available. And in preference to all of these is simply using Linux, but sometimes that's simply not an acceptable alternative to Windows (i.e. when you're making a product that you want to run on all popular OS's, supporting only Linux seems like a recipe for disaster. See Sage [2] for an example of an open-source project whose official line on Windows compatbility is "use Virtualbox.")
All your points (well, aside from mingw itself) are great ones. Cygwin is 100% a monster. But it's a monster that works the way I expect it to. MSYS is often lacking in things I consider basic that I miss going from OS X to Windows and trying to configure it is a super-pain-in-the-ass.
Colinux is actually pretty cool, but the need for an X server and the setup time is a pain in the ass.
Guilty of that here. There isn't much incentive really, other than perhaps getting Xmonad running, and it'll end up with some stuff breaking. I've got a nice bash shell, steam works and I can play portal, and it's good enough.
I certainly don't think it's a bad thing. Seems like without a MOC the programmer would have to write more code or file all that meta data in a separate file. So it's all good.
I was just curious since I haven't touched Qt in years and this post implies it is moving toward more standard C++.
Editor Domenic Denicola's website