[mattalex]

The problem was mostly that the only guy that was really backing the project (Christian Ude, SPD), was replaced with his successor (Dieter Reiter, SPD) who just didn't have the drive necessary to maintain the project.

The entire design of "LiMux" was doomed from the start: it was a highly customized version of Ubuntu that was only used in Munich (not even throughout the entire state). That made everything ridiculously expensive since the actual advantages of building on an open source solution was never realized. That is combined with the fact that "open source" and "cost savings" were used interchangeably when in reality the budget for Windows should have been pre-allocated into development, rather than cut.

The entire project was half-assed to begin with, which basically meant that Windows and Linux had to coexist since many crucial tools were never ported to Linux.

The "Microsoft killed it" story sounds realistic, but the truth is the much more boring incompetence in execution.

[bruce511]

I've no doubt that there was incompetence involved. Which is to be somewhat expected since Linux requires a set of skills to install and administer at scale, and few people have that particular skillset.

Equally I'm sure it was never going to be a cost saving exercise since firstly the cost of Windows is negligible, and second the conversion costs, in-house skill requirements, re-training of users, and porting existing software are all significant.

So you go from cheap software to free software, but from cheap IT staff to expensive, perhaps-incompetant, hard-to-replace IT staff.

If the savings aren't real then the fallback argument is privacy etc. But Libre Office runs fine on Windows. So by all means start there.

[sgift]

They always stated that it won't be cheaper. That was not the intention.

The main point was that Munich saw an opportunity to leave a monopoly for an option with competition. When the LiMux decision was made Munich had to change their whole IT landscape anyway (They ran on NT4 and even extended support was running out / getting extremely costly), so they thought: If we have to change anyway, why not change for something where we have more options?

[bruce511]

Having more options, in and of itself, has no value. Leaving a monopoly to go to an alternative option in itself has no value. Plus of course, since they decided to basically build their own option, they did neither.

Firstly, the number of options argument. Choosing between lots of options, over choosing the best option, isn't some intrinsic good. I don't support say football over Formula 1 because there are more teams. I choose one over the other based on which sport I prefer to watch.

Going the Linux route "because there are more distributions to choose from" is a dumb argument if Windows is the best OS for your organisation. [1].

Equally, you may decide that you'd prefer to foster competition. You might decide that it's worth subsidising competition just to have competition. You acknowledge you're making an inferior choice, but the are unaccounted upsides to others having that choice.

But govt adoption of Linux (especially creating your own distribution) does little to nurture competition. Munich didn't spend its budget improving Linux. They spent the budget implementing it.

There are reasons Windows and Office dominate the desktop, especially in large scale environments. You may not like MS, but that doesn't negate those reasons. If you make a choice to forge a different road because you don't like MS then you'd better be very clear on what those reasons are, and what challenges and costs you are likely to face. Because if you aren't, then earn those challenges and costs come up it's hard to wave them away.

I'm not saying all this to rag on Linux or OSS, or to say its impossible to run Linux etc. As per the GP, there are companies that have made a choice, that hire people with the right skills, that make it work.

But thinking that rolling out Linux at enterprise level is easy, or cheaper, or has better desktop options, or runs more bespoke software, or is easier to get good desktop software for , or is easier (cheaper) to buy hardware for, is wrong. And that wrongness will lead to project failure. And those failures don't help Linux, arguably they hurt it.