AFAIK the only reason stopping the Mixbus project to go Linux is a distribution problem, too many different packages for too many distros, and no unified easy one-click-run/install method OSX style for everything. Well, not until AppImage arrived! Have a look here:
Wouldn’t it be the perfect way to ship Mixbus for Linux? What could be easier than that?
I hope that some solution can be worked out, I didn’t realize this was so difficult to achieve. I emailed harrison a few weeks ago to let them know which distro I use, but that has changed since then and I now see how much of a difference between distros there is. I won’t bash anyone, but I am now on opensuse and it is running very well. If they release mixbus for a particular distro though, I will definitely use that for music as mixbus is the kind of program that will work exactly for my needs. Ardour with the same inline design and buses already set would do the same, although having that harrison dsp would be nice. I’ve been having great results with the linuxdsp plugs though, so if there is some way to set up ardour with those linuxdsp plugs inline, visible, and active, that would work for me.
This model (if done correctly) is very unpopular with distros. It results in multiple copies of what would otherwise be system libraries, with all the implications that has for security etc. etc. Its certainly a feasible solution if you only consider Mixbus, but it becomes problematic because of the way it totally circumvents almost everything that is good about the typical distro system.
The difficulty of producing a universal Linux binary (that will run on all versions of Linux) is undoubtedly a turn-off for budding new developers. I spent some time yesterday comparing AppImage with some of its rivals and I'd have to say that IMHO, AppImage does a pretty respectable job of solving this age-old Linux weakness.
I certainly agree with this assessment. It doesn’t however solve the big problem of system libraries vs included libraries, the latter of which is frowned upon by package distribution, however it also provides a much better support situation for commercial companies.
This being said, I do agree, the solution in general is a pretty respectable and decently thought out one. I wonder how difficult the FUSE requirement would make it for most distros though.
The difficulty of producing a universal Linux binary (that will run on all versions of Linux) is undoubtedly a turn-off for budding new developers. I spent some time yesterday comparing AppImage with some of its rivals and I’d have to say that IMHO, AppImage does a pretty respectable job of solving this age-old Linux weakness. AppImage isn’t perfect but it’s a lot more perfect than its rivals, most of which were little more than “repositories with a difference”. But with AppImage & Fuse, it’s clear that somebody has ‘thought outside the box’ and come up with a really quite elegant proposition.
I can understand why this approach is unpopular because, if abused, it paves the way for an unscrupulous dev to restrict licensing freedoms (something that would obviously be more difficult if he had to build binaries for hundreds of different distros)… But as long as it doesn’t get used as a ‘license abuser’ I think AppImage will be a worthwhile addition to Linux and could well be the magic bullet that really opens Linux to the masses (no disrespect to Mark Shuttleworht who’s done a pretty good job of that already).
The bottom line is that for Linux to become seriously popular, there are only two obvious routes:- the “universal binary” approach taken by AppImage - or the “one size fits all” approach taken by Mark Shuttleworth with Ubuntu. On balance, I’d argue that the universal binary approach is probably the lesser evil.
That’s true, you would end up why a system crowded with library duplicates and no security updates, but the good side of the typical distro system mostly applies only to free software that is distributable on distros repositories, while for a proprietary software with very limited impact on a system “health” like Mixbus, AppImage seems to me a viable solution…
I’m probably missing something, but don’t most of the improvements to MixBuss show up in Ardour anyway?
Most do, but that doesn’t mean Mixbus isn’t worth it. I have access to both on OS X and I use both. Which I use depends on what I am doing.