The second good news is that, in the case of Munich, the migration is financially self-sustaining. That's because the saved license costs exceed the costs of the project.
This is mostly because a large number of fairly standard desktops could be migrated early on.
What evidence do you have that the switch is financially self-sustaining based upon license fee savings? I'm not saying this isn't true, but I've never seen a reputable ROI analysis for Munich's switch. [caveat: my German is very very weak.]
I was in a hearing of a committee ("Enquete-Kommission Internet und digitale Gesellschaft") of the German Parliament (Bundestag) where the technical leader of the LiMux project mentioned the numbers in passing.
IIRC he mentioned that if you only consider the license costs, the balance is roughly even, and if you also consider that you need to upgrade the hardware less often (because of lower resource usage of limux vs. windows), limux comes out clearly ahead.
Sorry that I don't have anything to link to, and I notice that my source is biased too.
The hardware upgrade argument is interesting since the reason to upgrade hardware is because something breaks [same for Linux or Windows] or additional capability is needed [again the same for Linux or Windows] or just for the sake of upgrading [again the same].
It would seem that a hardware argument would be that Linux requires less expense for a replacement, e.g. Word Processing on a Raspberry Pi, though the idea of a Raspberry Pi in 2001 did not really exist. But the idea that Windows requires a more frequent upgrade cycle doesn't make sense since XP support only ended recently and still runs the hardware on which it was initially installed.
Anyway for me, the elephant for FOSS is support. There are just lots of moving parts in Linux and many development teams all rowing in different directions at best and losing interest in maintaining a package at worse, or perhaps being acquired by Oracle is worse. The price of all the innovation is the risk that comes from no roadmap.
> But the idea that Windows requires a more frequent upgrade cycle doesn't make sense since XP support only ended recently
But nobody knew that back then; iirc Windows XP EOL has been extended several times, and by the time the support ends, you should have migrated all of the remaining machines. So planning an upgrade for 2013 or 2014 would have been downright irresponsible.
> Anyway for me, the elephant for FOSS is support. There are just lots of moving parts in Linux and many development teams all rowing in different directions at best and losing interest in maintaining a package at worse, or perhaps being acquired by Oracle is worse. The price of all the innovation is the risk that comes from no roadmap.
There are lots of paid support contracts for open source software. Also you can mitigate much of the risk by choosing projects from healthy communities with high bus numbers.
When the company from which you bought support goes out of business, at least for open source projects you have the chance to find another company that provides support; good luck with that when the source is closed.
> But the idea that Windows requires a more frequent upgrade cycle doesn't make sense since XP support only ended recently and still runs the hardware on which it was initially installed.
Windows is more and more a resource hog with every version update. Many machines able to run a WinXP are not able to run a Win7 or even 8.1, same for Vista and 7/8.1.
Windows 7 had lower system requirements than Windows Vista. Windows 8 and 8.1 have the same system requirements as Windows 7, and in my experience often use less resources (presumably as part of their focus on mobile/tablet scenarios). The trend for the last 8 years has been one of reduced requirements, not increased.
That is the usual commercial bullshit, but when you have a recent computer bought with windows 7 that can not upgrade to windows 8, you understand your pain.
But upgrading from XP hasn't been necessary until last month. Until then, any machine that shipped with XP could be kept up to date via Download Tuesday without any change to the hardware, and upgrading remained as much a choice for Windows XP as for Linux.
It's like saying that the existence of the Linux x64 kernel requires upgrading 32 bit machines to x64 hardware.
This is mostly because a large number of fairly standard desktops could be migrated early on.