Munich moving to Windows 10?

[GreyGeek]

I just saw an article in TechRepublic which claimed that Munich was considering switching from Linux to Windows 10. It struck me as being very familiar to similar articles published about the time that Munich completed the move FROM Windows to Linux, begun in 2004 and completed in 2013, for 15,000 city employees. (That ignores the question of why a city of 2.6 Million inhabitants would need one city employee for every 173 citizens?)

Investigating, I found that the same person who caused this stink in 2014, then newly elected mayor Dieter Reiter, a Windows man, is the same one making the stink all over again this year. And using the same arguments. The basis of it is an unverified Microsoft report claiming that it cost Munich over $60M, not $10M to move to Linux. The report is bogus, of course, as is all of Microsoft's propaganda. On the contrary, Munich is searching for Linux Administrators. The job opening was posted on Monday, November 7,2016, two days ago! (Google Translate probably does not do it justice!)

You are a Linux professional and are looking for a new professional challenge in which you can expand your knowledge from A like Awk over L as log file analyzes to T like troubleshooting in SUSE, Debian, Ubuntu & Co and use it productively? Then go with us the next step! Use our personal contacts to employers nationwide! We recommend you to our customers and thus facilitate your entry into the company! On behalf of our well-known customer company in Munchen, we are looking for a dedicated and reliable Linux Administrator (m / w). Good work life balance as well as an indefinite employment contract. Your tasks - Design, planning and installation of Linux systems - Troubleshooting (2nd and 3rd level support) - Analysis of system messages and log files - Further development of internal IT infrastructure - Various project processing Your profile Successful completion of computer science or comparable technical training First job experience (SUSE, Debian, Ubuntu etc.) Knowledge of the script languages Bash, Awk, Perl Very good knowledge of German as well as good English skills You are interested in this job? Please apply online under reference number 135806. If you have any questions, please do not hesitate to contact Alexander Simon of Robert Half Technology.

A similar article asks an important question by Matthias Kirschner, president of the Free Software Foundation Europe (FSFE):

Kirschner also questioned why Accenture was chosen to co-author a report assessing the use of Microsoft software, when the consultancy runs a joint venture with Microsoft called Avanade, which helps businesses implement Microsoft technologies.

Accenture, for folks who do not remember the debacle of the 2008 London Stock Exchange crash that kept them off line for an entire day (the second crash), was the company that co-authored with Microsoft the .NET "solution" for a stock trading system which was supposed to achieve a transaction time of under 2 milliseconds, IIRC. Their dreams came to a crashing end when the "mature and secure" .NET caused a Billion $ crash on the London stock market trying to reach their promised transaction speed. The LSE ended up replacing .NET and Windows with a Linux stock market program which for five years had been running five times faster than Microsoft's unachievable speed goal. The LSE bought the company which wrote the program. The Blue Screen of Death on the ceiling of the Bird Nest at the Chinese Olympics, a SilverLight failure which made it clear that MoonLight could not be any better, hastened .NET's fall from grace and MONO's evaporation. Mono was really a "kill Linux distros" plan, IMO, because while the CLI components had been put under Microsoft's version of an open source and a letter promising not to sue, the important GUI components were not. MONO enthusiasts denied, papered over or ignored that important caveat. They might have succeeded because of the media juggernaut fueled by Microsoft money, but then the LSE crash occurred and that was quickly followed by proof that the claim in Microsoft's ad, "The Reliable Times", in which MS claimed that the LSE chose the .NET solution over Linux after a rigorous competition, was a lie.

So, the Munich story is just Microsoft up to its old tricks. The convicted felon has not lost its stripes, and Munich is not switching back to Windows.

[wizard10000]

I have actually done a formal cost benefit analysis for a proposed move to Linux for a rather large government agency and sad to say, the sticker shock at cost of migration kept the rather large government agency on Windows.

Software costs are inconsequential. The average large business spends about 12% of its IT budget on software and long-term there is a cost savings to move to Linux but the short- and medium-term cost of migration for a large enterprise is horrendous. In this particular agency the cost of migrating 3,500 workstations, training about the same number of users to Linux plus increased service desk and deskside staffing plus recoding Windows applications to run under Linux made the three-year cost of migration several million bucks.

If you start with Linux out of the gate you save money - but the short term cost of migration for a large enterprise is unbelievable. The cost of migration back to Windows would also be ridiculously expensive

[GreyGeek]

About the time Munich began moving its 15,000 city employees to Linux Microsoft fought the trend by secretly funding "migration studies" (we learned about them in the Combs v Microsoft Trial, and about O'Hara and Diddio's paid propaganda) which purported to show that it was more expensive to migrate to Linux than "stay with Windows". Their technique was to swamp the software costs with those of hardware and IT suport ("maintenance"), making the software costs appear to be insignificant. However, it is obvious that hardware costs are software agnostic and thus a wash, so it boils down to maintenance and software costs. WIndows maintenance costs (usually centered around recovery after malware infection, or buggy software crashes) was more expensive because Linux was more stable and immune. Linux was free. Eliminating the redundant costs Linux was cheaper and didn't have expensive upgrade treadmill costs forced on the users.

I, too, was involved with the move of a gov agency from Windows/NetWare to Linux and open source software about a decade ago.

I found that the hardware turns over about every three or four years because to stay with older, slower hardware is to waste the time of employees (13,000), which translates into LOTS of $$$. The new hardware can be naked and only one copy of Linux is needed to install Linux on all of them, and that copy is free. Installation costs are a wash because it takes about the same amount of IT time to install Linux or WinX, which is usually a batch process. Update costs for Linux are only IT time and are also a wash for the same reasons, except that CAL fees don't have to be paid to update Linux, so Linux wins that one. Software is free for Linux but license costs are significant for Windows software unless they install open source software on Windows! Microsoft is constantly churning the upgrade treadmill to squeeze more cash out of clients. Linux doesn't have a cash churning treadmill.

Everyone talks about the big cost of user training when moving to Linux. That is essentially malarkey. No one was born knowing how to run Office or OutLook. That many learned to use Office in HS or College speaks only to Microsoft's hegemony. Every time Microsoft churns the treadmill with a new version of Office users have to learn again how to use the new version, and pay more money for the privilege. Microsoft also FORCED upgrades by not including the old file formats in the new version. Those using the older version of Office, for example, could not read docs written in the new doc format, and the newer version could not save a doc using the old format, FORCING users of older versions to upgrade. $$$

Using LibreOffice is no more difficult than using Office. Both are run using the mouse and keyboard, and have very similar structures and help menus.

The big problem are the suits (management) who usually know next to nothing about software or computers and whose primary role is to protect their own job (in gov agencies). So, when choosing between Oracle and PostgreSQL (in my experience) they chose Oracle because "it has support". Showing them the free PostgreSQL support sites didn't matter because "anything that's free can't be worth much". Such was their Neanderthal thinking. Oracle started out expensive ... one license per server or blade. When servers/blades began to carry more than one core Oracle switched to one license per core. In our case that multiplied the cost by 2X, then 4x and then 8x times. With virtualization the license costs increased even more. What used to cost around $100,000/yr exploded to over a million/yr. Shot the budget. That doesn't take into account the upgrade treadmill, with each new version costing more than the last, with previous versions falling to EOL non-support. The support? Post a ticket and expect to get a reply in a week, if you get one at all. Our Oracle admin ended up getting rapid response support, and free, from an online support site staffed and visited by other Oracle users. So the support money is wasted but you can't get Oracle without paying for it.

The migration came to a sudden end when a new governor appointed a new tax commissioner, who hired an assistant tax commissioner. The governor's "economic improvements" promised during the campaign was to replace the single tax commissioner with two. The assistant TC was a suit that only knew how to use Excel and Word. She immediately dictated an all Microsoft shop. Costs exploded. Productivity cut in half, as did the network speed. Malware infections became so bad that IT was forced to install a Linux Internet gateway with AV software. Sometimes you don't get what you paid for.

[wizard10000]

...Everyone talks about the big cost of user training when moving to Linux. That is essentially malarkey.

Respectfully disagree here. I've done end user support professionally for > 30 years and right now am managing internal IT service desk support for a 10k user gas and electric utility. If you make something different people call the helpdesk; and I do helpdesk fairly inexpensively (about $14 per call).

Migrate 10k users to Linux and at least a quarter of them will call the helpdesk at least once during the month, so that's $28k in additional service desk cost alone (at least for the first month - subsequent months will be a little better). Service desk and deskside resources do need to be trained to support the new OS and for planning purposes we set the cost of identifying, archiving, reimaging and restoring an end user's data at four hours per workstation. Cost of education is pretty far from insignificant. Recoding proprietary applications to run under Linux isn't an insignificant cost either

Also, not all users use computers - about half of our users climb power poles or work in generating plants for a living and retraining field workers is a complicated and expensive process. Right now an untrained user can generally add a network printer on a Windows machine; the chances of that same user being able to install a printer driver under Linux is just gonna generate another call to the service desk.

As mentioned, the cost savings are there in the long term but for the first three years or so someone would be hard-pressed to show a cost benefit.

Think we're gonna have to agree to disagree, GG

cheers -

[GreyGeek]

....

Think we're gonna have to agree to disagree, GG

cheers -

I guess so.
However, in my experience help desk calls are independent of the OS used and more a function of the intelligence of the user. Those who required help with Linux had required about an equal amount of help with Windows. I've been retired more than 8 years, and my experience was around 2002-2005. A LOT of things have changed since then, including the availability of commercial Linux software solutions. Regardless, it costs no more to write in-house apps for Linux than it does for Windows, and I've done both for more than a decade.

Cheers back to you, Wiz!