OSNews

It doesn't require anecdotal evidence.

Porting from Windows to two Linux distros costs more than porting to OSX, especially if the application uses sound or video. On that basis alone you can argue that fragmentation holds desktop Linux back. Keep in mind that porting for a commercial ISV include testing and support, which is further increased by distros like Ubuntu that are constantly releasing major updates.

Linux is a PITA for commercial developers and it has been that way for years. Distro differences are mitigated by releasing the source and then having package maintainers handle the porting and distribution. Once you step outside this model the costs go up.

You can't argue that fragmentation is a net positive or neutral aspect for commercial developers when there is clearly an additional cost involved. "

Yes, I do agree with the answer.

Just consider another example: packaging the same program for 20 distros waste 20 times time needed for just one distro.

Getting a specific version of an app packaged for your distro is almost impossible (so you have to use the version shipped with the distro while you could use a more recent one, if someone packaged it)

Fixing problems and helping other users supporting them is time consuming because you need to know the "internals" of every distro on the planet, which is impossibile.

Deploy working solutions requires you to test on 200zillion distro. Or at least you have to develop in the same environment where you are going to deploy (which could be hard or impossible if you need to work for several customer at the same time)

Non-professional user may not be able to use another user's Linux machine because it may act / feel completely different from its own setup.

We can go on forever with these scenarios.

So, and I'm writing to the one who says that fragmentation is not bad, what are your evidence that is is good? (or at least, neutral?)

I actually worked on a .NET project where Linux was considered, but only because of OSX.

OSX ended up being dropped because Mono runs through X11 and the application would have required a lot of work to make it look acceptable. Monomac at least builds on Cocoa but is a work in progress. Then on top of it Mono is behind compared to the real .NET which amounts to it being unappealing for any company that wants to make a great looking Win7/Vista application. It's better suited for in-house porting.

Mono applications look fine in Linux but a divided 1% is not an appealing platform to target, especially since there is no commitment to provide a stable platform for commercial companies.

It needs to be remembered that while a port to either Linux or OSX can be profitable the investment is typically weighed against a range of options including further investment into Windows or even web or mobile platforms.

What needs to happen is for Qt to get its remaining OSX issues resolved which will lead to more cross-platform development. When applications are designed to be cross-platform from the beginning the cost of porting to additional platforms is significantly reduced.

Mono haters should be happy to know that Windows devs are not gravitating towards it. The move to WPF has left it floating in the water. Mono bashing is not even needed. Qt will hit a tipping point and no one will care about Mono.

First "Porting to two OSes costs more than porting to one." - Is anyone even surprised? Second, I was talking about (and we were discussing) Desktop Linux fragmentation in the form of GNOME vs KDE, not distro fragmentation.


Pick a version target and stick with it. This is why enterprise distros and Debian are good. Again, don't change the subject: Multiple DEs are not the problem here.


Which has exactly what to do with a choice of Desktop Environment? Pick a distribution and target it, pick RHEL if you want a stable target.


All I have to show is that fragmentation is not a net negative, or that it cannot be proven to be a net negative. But I don't even need to do that: I am not arguing against various platform problems surrounding Linux in general, I am arguing about whether GNOME vs KDE vs GNOME+Unity fragmentation is hurting Desktop Linux. It isn't. You still haven't even attempted to describe how it might.

That works for a database company targeting a conservative distro like RHEL but a gaming company needs to support the latest version of Ubuntu.


It makes Linux less appealing to ISVs since multiple environments increase development and support costs. Proper Qt integration within Gnome is really what is needed.