Does VirtualBox VM Have Much A Future Left?

[duby229]

The way I read the comments here about CLI versus GUI is this: documentation.

Yep, that "big word" that most developers don't seem to bother with. Most developers do a decent job writing code that works, but then they leave the documentation up to whatever time they have left, or have someone else, like a volunteer, write the dox. Either case: FAIL.

Manpage style documentation seems more targeted at users that already have a good knowledge of the program, but they just need that hint, that clue, that "jog to their memory" to help them figure out some arcane CLI option. Even a lot of online documentation is little more than Manpages sucked into webpages and possibly reformatted.

Then there are users that just like working in a GUI. In that case, documentation should also be written for the GUI format. It does not need to be fancy, "skinnable", or even support all your Facebook/Twitter/Instragram buttons. It just needs to be helpful and easy to use.

Good documentation to me clearly explains what various options mean & do, not to mention any applicable "default values". When documentation is in it's GUI form, a nice screenshot or two is always helpful, like the old Windows ".CHM" files could support. An index is needed for any program that might have more than 2 options. Having various "worked examples", especially in CLI for CLI users, really helps develop an understanding of how to use a program.

Alas! Documentation is usually an "after thought".

It's obvious you've had issues in the past with poor documentation. That's actually where I spend a lot of my time, there are several projects that I contribute documentation and bug tracking for.

I see folks say over and over again, "If you don't like it, then fix it", or "if you don't code, then you don't have a say so", but there are other ways to contribute too. In -a lot- of cases it's documentation that needs the most work. As a "non-programmer" that just happens to be one of the areas I spend a lot of time contributing to.

[DeepDayze]

VMWare player isn't OSS, but if you want something that works across linux and windows it's better than vbox. At least VMWare cares about their product. Oracle obviously doesn't give a shit about any of Sun's products.

VMWare is the most commonly used virtualization software in the enterprise so if you use VMWare at work you would be comfortable for using at home

[duby229]

VMWare is the most commonly used virtualization software in the enterprise so if you use VMWare at work you would be comfortable for using at home

If you need it at work, you would have to pay for VMWare Workstation. It's reasonably priced, especially if you consider that your paycheck comes in part because of it. At home you can just use VMWare Player.

[dschaefer]

qemu?

Correct me if I'm wrong, but isn't VirtualBox already a fork of qemu? I wonder if it makes sense to bring some of the great things about VirtualBox, like it's 3D support over to it. There's also the Android emulator out there, another fork of qemu, which also has some support for 3D.

I would just think focusing the OSS community on qemu, which already has a lot of it, would be the right thing to do. And I wonder why it hasn't already.

[SystemCrasher]

I'm using KVM these days. Because unlike VBox its kernel module comes by default with most distro kernels. So just install qemu (and some GUI if you need or want one) - done! On other hand, VBox requires custom out of tree kernel module and there're inherently some pitfalls. Then Oracle is known to be greedy to degree it happens to be counterproductive. Hey, why USB passthrough should be proprietary? Okay, KVM virtualisation does not haves this idiocy attached. Then, kvm virtio is quite developed, reasonably fast, and there're ton of settings for all occasions. Best of all, it can be controlled from command line if I need automation, etc.

So what is the point to use VirtualBox these days? Has it got any real advantages over others? Or why someone supposes I would mess with foreign kernel module?

[Calinou]

Most of the Vbox software IS opensource, so anyone could takeover the development and advance VBox.

[overwhelming sound of crickets]

VirtualBox has issues preventing it from becoming fully free/libre software:

* a proprietary compiler is required to build the BIOS,
* proprietary guest addons and extensions are recommeneded.

However, it looks like VirtualBox is the only available, easy-to-use cross-platform solution that isn't VMWare. It also features some 3D acceleration (although that is less useful these days thanks to LLVMpipe), unlike KVM/QEMU.

[Nelson]

I'm using KVM these days. Because unlike VBox its kernel module comes by default with most distro kernels. So just install qemu (and some GUI if you need or want one) - done! On other hand, VBox requires custom out of tree kernel module and there're inherently some pitfalls. Then Oracle is known to be greedy to degree it happens to be counterproductive. Hey, why USB passthrough should be proprietary? Okay, KVM virtualisation does not haves this idiocy attached. Then, kvm virtio is quite developed, reasonably fast, and there're ton of settings for all occasions. Best of all, it can be controlled from command line if I need automation, etc.

So what is the point to use VirtualBox these days? Has it got any real advantages over others? Or why someone supposes I would mess with foreign kernel module?

VirtualBox has a fairly complete API for programmatically creating environments. libvirt nicely replaces most of that though. It doesn't seem like it should need a ton of work, it's always been a little rough, IMO, it's heavier than VMWare.

Are people leaving it because it's not updating every 3 weeks? Or is it not updating every 3 weeks because fewer people use it?

[Akka]

IMO that is the wrong attitude. OSS devs shouldn't have to work for free. They -should- get paid for their work, they deserve to get paid. A fork of that size is bad for everyone. It duplicates effort that could be going into a different existing project.

Forks are not always the answer. In this case, without commercial backing a fork would have no chance for success.

Its not going to happen if no one know how they should make money on the product. A company is not going to pay for the development if they don't is able to make money of it in some way. Both Oracle and Sun obviously failed with that. And the plausible result is the slow death of virtualbox.

[michael-vb]

As one of the four developers...

...I think there must be some confusion here, there are definitely more than four of us! I think reports of our death must have been slightly, should I say, exaggerated? I won't go into too any detail, but as one commenter said, take a look at our svn log to see what we are working on. Hope you will enjoy the results!

[darkbasic]

I won't go into too any detail

Why not? Lots of peoples here are pretty interested in what you are working on for the next release.