Just start your presentation with the line "And then this one time, at band camp..." and you'll have their undivided attention.

Training people to use Linux is pretty simple unless they're dense.

Dense, no, they are government workers.

When they decided, their best information AT THE TIME was that Linux was the more expensive, yet preferable option.

If they simply looked at cost (which they didn't) they would have gone with MS.

Hindsight is only available after the fact, I was commenting on the inputs they had to their original decision. That they would ultimately save money was, at best, a leap of faith when they committed to Linux over MS WinXP. They saw (and have since realized) many reasons to choose Linux and did.

Because Linus referred to it as Linux when he released his kernel, and when other people added a large number of GNU utilities to that kernel and called it an OS they simply perpetuated the name.

Any thoughts of a greater "conspiracy" is a wasted effort - maybe if RMS had actually focused on writing his own kernel instead of taking a decade to decide on the "proper" kernel his suite of software utilities and another decade to write the kernel then it would be regarded as something more than a set of tools added to Linus's OS.

I always laugh at those that say we need to "read the source code" - with apologies to Rep. Conyers.

You imagine governments are staffed with computer professionals capable and motivated to worry about their OS being compiled by the Mfg.?

At what level of government do you imagine would be performing code reviews and building their own OS images? Federal? State? County? Municipal? Here's a better idea, keep government worker desktops off the Internet, then it doesn't really matter how vulnerable a desktop OS is if there is an air-gap between it and the internet.

Yeah, why would MS try and retain clients? They'll just let cities drift off into "roll your own" land and watch their business revenue shrink...

Any large-scale deployment takes significant man-hours to achieve, but can be made easier through the use of imaging and common platforms. If I standardize on only a handful of models of computers then I can load-up the OS and build everything that I need for that OS on each model, then simply duplicate the drive over all of the others of that model, change the few things that need to be changed (name, network credentials, possibly some security hashes) and I'm done.

It's called WDS, and it's included for free with Windows. Clonzilla, Ghost, and other tools work equally well with MS and FOSS system images.

This is arguably even easier in Linux than in Windows because there are no particular licensing issues with just copying a Linux installation or with how many Linux installations are deployed. One's backend servers are now for updating and package management rather than for licensing.

How hard do you imagine MS software licensing is? You configure one server VM to serve out licenses, and when new license codes are available, the admin simply adds them to the license repository. The client OS and applications (MS Office) are pre-configured to seek out a KMS license server. Once the server is configured, there is no need to even think about licenses on client machines, it just works.

And with Microsoft deciding to change their UI every few years now, coupled with competing UIs from Apple and Google, it's much easier to change people to a diffrent platform when they have to learn a new UI anyway. Had Microsoft kept variants of the Windows 95 UI going past Windows 7 then it would be harder, but with the Metro debacle it's a lot easier to make that change, and since most users won't go deeper than the UI anyway it's not so bad.

And by "every few years" you mean every decade? As you alluded to, Windows 95, Vista, and Windows 7 have essentially the same UI, conversely, Ubuntu has changed it's desktop interface more frequently.

The hardest part is training the support staff if they've been Windows-centric their whole careers. Somehow just reiterating that everything-is-a-file isn't enough, and many professionals struggle to understand UNIX-style paths.

Yeah, because users that have learned "to the click" to work in Office 2010, 2011 (Mac) or 2013 will have virtually no learning curve under any of the free Office alternatives...

You ignorance of the Windows ecosystem makes it easy to find fault in it - you can simply say you prefer Linux and leave it at that, but your arguments against Windows are really rather trivial issues, nothing more.

Canonical offers a comprehensive management suite for desktops and servers, that in may ways compares with Windows AD and associated tools. Canonical charges about $200-250/system per year (I assume volume discounts are available, but I'm not privy to them), while annual software license costs for most MS software users is well under that number (for example, schools can get client OS license, MS Office, server CALs, and misc other MS software for $35/desktop per year).

There are other options, including "roll-your-own", but when considering 15,000 desktops the task can become overwhelming and take a number of years to fully design and implement, and what to do during that transition period?

Support of older hardware is a meaningless metric, will the city of Munich be purposefully running older hardware bought surplus/off-lease, or will they buy current hardware going forward? Systems have a certain useful life, and buying machines mid-way through their useful life, while extremely cost-effective, can result in more frequent hardware swaps/upgrades, increasing labor costs but each iteration will cost less.

Put simply, let's say a given laptop has a five-year useful life, buying a laptop that is three years old doesn't extend the useful life of the laptop out to eight years, you are instead buying the last 2-3 years of it's useful life.

Sometimes organizational goals don't match organizational realities.

Did Munich buy 15,000 identical desktops & laptops for all users, and will perform similar massive (government-wide) forklift upgrades going forward, or will new models be brought in over time, creating an ever-changing mix of systems?

My corporate IT background tells me the latter is more likely, but hey, maybe Munich is different.

More money, but to local people.

They were faced with a "massive" migration to either WinXP or Linux, on a cost-basis, MS was cheaper - functionality-wise, benefit to the community Linux was superior, and they choose Linux.

I didn't judge the decision, I simply reported what was written in the article. Personally, I think they made an excellent choice by keeping the money local, even if it was greater than the foreign (MS) option.

I discussed their decision, and when they made their decision Linux was the more expensive option and they took it.

Save money and go with MS?

What saved money? They went with Linux despite it costing more than the MS alternative - it was buried in the fourth paragraph of the linked-to article.

Rather than send fewer dollars to the US, they spent more dollars and hired local Munich companies to handle the migration.