The Munich Revolution

Lately, in the news, we read that agitators are still pushing to roll back GNU/Linux desktops. It turns out that GNU/Linux is thriving in Munich. A recent update has these highlights:

  • The complement of PCs has risen to 24K
  • 18K run GNU/Linux (75%)
  • Most users don’t care about the OS and most were not even aware of the OS in operation. They used OpenOffice.org/LibreOffice on both OS so they don’t use the OS much, just the applications. One complainer about GNU/Linux was actually using XP…
  • Munich maintains about 4K software packages for their system, so they do care that the software is FLOSS.

See a report from DebConf 15

IMHO, they are still doing things the hard way. They have a lot of bugs to squash because they depend on Ubuntu. Life would be easier sticking with Debian GNU/Linux. Their release-cycle is about 3 years so they get only two years’ use of Ubuntu LTS. They would be much better off using Debian who take about 2 years getting the bugs out and releasing LTS good for 5 years. They could do better by having the non-mobile PCs run as thin clients. Then fewer machines would need to be maintained. After that their biggest problem is that they still have to support That Other OS which is a real dog.

Still, they are much further ahead not paying M$ for use of their own hardware and enduring endless re-re-reboots and malware.

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology and tagged , , , , , , , , , , . Bookmark the permalink.

38 Responses to The Munich Revolution

  1. oiaohm says:

    Something else to remember even that number of non Linux machines has doubled in Munich there are still the license cost savings of deploying Libreoffice instead of MS Office on most of those machines. 80 percent cost saving remains even that the Linux percentage has dropped.

  2. oiaohm says:

    Robert Pogson you are so right Drloser complete did not understand what I said.

    French Police is more the model I would be looking at for Munich. French Police is Linux Thick Client with Windows provided by Thin Client.

    Basically there are two successful models conversions that have been running for over 10 years by different groups that archive 90% license reduction.

    –Do tell us how your 20/20 hindsight plan would have produced a viable 90% Linux thin-client solution for Munich, albeit about ten years after the original plan.–
    I do not need 20/20 hindsight. In Munich everything is thick client if it be Windows or if it be Linux. All I need to-do is compare Munich to those who have achieved 90%+ and see what the difference is.

    The lack of a Windows thin-client system is the difference between Munich and the ones that have achieved 90%+

    French Police have using 8 different solutions to provide Windows by Thin-client. Included was Windows Terminal Servers early on.

    DrLoser the French Police had crossed 80% Linux before Munich had deployed there first Linux machine. Munich has decided for some reason not to follow the French Police model. Or maybe they are they are just not in any rush. Munich is a horible model to consider replicating.

    The main reason why the French Police is thick Linux is that operations have to be able to proceed even if central servers are no more.

    The reality here is those with Windows Thin Client can deploy far more aggressively as well. There are documented Migrations like French Police that in 7 days goes from 99 percent Windows to 80 percent+ Linux.

    French embassies are a good example of Migrations that don’t work. Each of the embassies had on average 1 computer. Converting those to Linux with no access to Windows applications was dead so fast it was not funny.

    For home users until we can successfully run Linux and Windows at the same time at full speed them doing a mass migration is unlikely. But with Intel and AMD both working on hardware support for doing just that the day of the Windows only computer might be numbered.

  3. DrLoser, putting words in others’ mouths, wrote, “what you are saying here is that the whole concept was an utter disaster and an expensive farce, not because of the choice of OS vendors (M$ vs FLOSS), but because they didn’t go full-bore thin client?”

    That’s clearly not what oiaohm and others have written. Munich did multiple things: a complete revamp of IT, a switch of a large proportion of clients to GNU/Linux, and a rationalization of applications, minimizing costs, duplication etc. I think they would have been better off with a thin client solution as the end goal but they only used thin clients as a means of virtualization of applications, particularly That Other OS v GNU/Linux applications. Largo, FL, for instance used thin clients widely and made huge savings, cutting down capital cost of the client suite and operating costs of the applications while optimizing spending on servers. There was little downside to using thin clients. I doubt Munich is a hotbed of full-video creation or whatever barrier there might have been to thin clients everywhere. A clerk using a browser, LibreOffice and some database just doesn’t need a thick client. That is one savings Munich could have made and they may make it in the future but complaining they messed up by not doing that is silly. It is very easy to switch from thick to thin. I’ve done it a lot and there was only one tiny problem with OpenOffice.org back in the early days, not unlike the few problems thick clients had. Today, the major point about not using thin clients everywhere is that many websites really load the client down, e.g. 1.5gB RAM just to open a few pages. When Munich made its move, a 64MB thin client was very feasible and folks budgetted just 100-200MB RAM on servers. For Munich’s own web-applications, there likely would have been no problem whatsoever.

    Who knows why Munich did not use thin clients widely? Perhaps even they don’t know. Perhaps they just wanted to get on with the obvious plan of changing OS and applications. Many schools and banks can do a changeover on a weekend to switch to GNU/Linux on thin clients.

  4. DrLoser says:

    Reality 90% goal by Munich with no plans to deploy any thin-client solutions was dreaming.

    Fascinating, Fifi. Do tell us how your 20/20 hindsight plan would have produced a viable 90% Linux thin-client solution for Munich, albeit about ten years after the original plan.

    So, what you are saying here is that the whole concept was an utter disaster and an expensive farce, not because of the choice of OS vendors (M$ vs FLOSS), but because they didn’t go full-bore thin client?

    Interesting. Let’s get back on a level playing field here. If you insist upon a thin-client solution, then you have two basic alternatives (I’m going to leave out the possibility of Solaris, although in 2004 that was quite plausible).

    1. M$ servers with Citrix or RDE or various kiosk-style alternatives.
    2. Red Hat servers with … some sort of thin client, it doesn’t really matter what. Except that you have to choose between dedicated thin clients of the sort that are used in Insurance Companies and the like (aka glorified 3270 terminals), or repurposed desktops. And in the latter case you have to choose the thin client that works for Munich … which apparently never occurred to the neck-beards who borked Munich in the first place.

    Now, I’m not saying it couldn’t have been done, Fifi. I’m not even saying it shouldn’t have been done.

    I am, however, stating the obvious: it wasn’t done. Two questions, therefore:

    1. Why wasn’t it done?
    2. How would you have done it, Fifi?

  5. Deaf Spy says:

    Very naughty boys, these Germans. Have I pointed out that they are Catholics?
    They started two world wars. And almost won them.

  6. oiaohm says:

    Reality 90% goal by Munich with no plans to deploy any thin-client solutions was dreaming.

    http://www.amd.com/Documents/Multiuser-GPU-Datasheet.pdf
    This is the second todo hardware Multiuser GPU. Intel GVT-g was first.

    The problem what stuffs you with thick client only solution that is Munich is a employee who job is 95 percent Linux 5 percent windows. Of course the employee will not want to go to another desk. Yes reality the Windows License by thin-client access could be shared between 10+ users. This is what allows the 90%+ Linux to happen.

    Like with word documents you only need to use MS Word if the document does not open in Libreoffice. You can give more staff access to Photoshop and the like as well.

    Reality you have paid for software you want to see it in 100 percent usage. A single staff member is never going to achieve that.

    Munich is not a super great example.

  7. oiaohm says:

    –In October 2013, the city of Munich had migrated over 15,000 desktop PCs (of about 18,000 desktops) —
    This is from the wikipedia and it cited. So 2013 3000~ Windows and what ever else.

    Today Linux increase 3000~ to 18000 and Windows/OS X/whatever has also increased by 3000~ to 6000~.

    It 6 thousand increase in number of PC not 8 thousand as you said DrLoser. Half of that number Linux. Half is a mix of other stuff.

    At peak Munich was only 83 1/3 percent. DrLoser goal was 90% but Munich never achieved that. So it gone back by about 8 percent but this is nothing abnormal.

    http://davelargo.blogspot.com.au/ Largo and French Police did achieve 90%+ If you are after 90%+ migration you don’t look at Munich as a model.

    Munich is not giving us details on what the mix of other OS is in the 6000.

    DrLoser as normal you barely got a figure correct. Come on how hard would to been to have checked the wikipedia first.

  8. DrLoser says:

    One small extra observation, Robert.

    â—¾The complement of PCs has risen to 24K
    â—¾18K run GNU/Linux (75%)

    Correct me if I am wrong, but circa June 2013 or so, the figure for Linux (obscure and not generally available to the plebs, but LiMux is still Linux) desktops in Munich was something like 90%, as widely advertised. From memory the number of desktops involved was something like 16K.

    Do you have a theory as to why the ~8K new PCs run at approximately 50/50 LiMux to Windows?
    Shudder.
    Never mind, the web-page views probably make up the difference.
    Web page views is basically the main motivator for a large German municipality to acquire 8K new desktops and laptops.

    Very naughty boys, these Germans. Have I pointed out that they are Catholics?

  9. oiaohm wrote, “Robert Pogson bugs are still being fixed in JFS. “

    Yes, but the bugs are not even functional:“Not much this time. Just a one-liner format fix”.

  10. oiaohm says:

    https://wiki.debian.org/DebianLenny
    2006-11-16 is when Debian Lenny starts. This is when you QA teams start with it testing and reporting issues.
    Then it has 3 years end in 2012 in production.

    Newer releases have 5 years in production release instead of 3. So overall support time will be closer to 8 years than 5.

    https://wiki.debian.org/LTS/Using

    But of course DrLoser you do need the read the documentation on what is covered. Of course you are too lazy to have done this instead you want Robert to repeat to you what is in the Debian wiki you never read before coming saying crap.

  11. DrLoser says:

    ” What, precisely, do you mean by “support?” ”

    Psychological support to help vict… users cope?

    Sounds about right, Kurks.
    Anyhow, Robert, about this five year long Contractual Goodness of Debian Lenny.
    Where did it all go wrong?

  12. oiaohm says:

    https://gitlab.com/veo-labs/linux/commit/78d5dcda92a17f17132671c269ea2c3a17688649
    Robert Pogson bugs are still being fixed in JFS.

    JFS 4.2 now moves it default git repository off Kernel.org hosting to github.
    https://github.com/konis nilfs maintainer.
    https://github.com/kleikamp

    Just because the maintainer-ship slot is still held by an old team and absolutely critical patch work is being done(the stuff where the code refuses to build) does not make me trust it.

    Dave Kleikamp over all kernel and OSS activity has dropped since he moved over to Oracle.

    Basically JFS is on my list not to use with option to change mind if new Maintainer takes over or Dave Kleikamp/those submitting patches to JFS activity increases.

    JFS is down to one party doing only the patches need so it builds.

    JFS Rename fault in fixed in kenrel 4.2 was found because it failed default kernels static testing. Other file systems like the XFS one you reported is because someone is running pressure tests. IBM is one of the few places that has the resources to pressure test file systems.

    Popularity contest is no sign of quality. There are quite a few poorly maintained to unmaintained that rate quite highly in Debian popularity competition.

    I have been caught in the past trusting a part of Linux where the maintainer is asleep at wheel.

  13. oiaohm wrote, “The fact over 12 months has past since the prior team and a new team has not taken over I no longer trust JFS.”

    Shaggy still works on it. I know he’s now with Oracle, but…

    This is from the MAINTAINERS file in linux-4.1:
    JFS FILESYSTEM
    M: Dave Kleikamp
    L: jfs-discussion@lists.sourceforge.net
    W: http://jfs.sourceforge.net/
    T: git git://git.kernel.org/pub/scm/linux/kernel/git/shaggy/jfs-2.6.git
    S: Maintained
    F: Documentation/filesystems/jfs.txt
    F: fs/jfs/

    It’s “maintained” as far as Linus knows. Shaggy’s still active on LKML. I don’t believe something has to be constantly changing to be relevant. Sometimes stability matters. The file-system is one of those times.

    Then, there’s things like this:“this is a second version of the patches to fix data corruption in mmapped data when blocksize < pagesize as tested by xfstests generic/030 test. The patchset fixes XFS and ext4. I've checked and btrfs doesn't need fixing because it doesn't support blocksize < pagesize. If that's ever going to change btrfs will likely need a similar treatment. ocfs2, ext2, ext3 are OK since they happily allocate blocks during writeback. For other filesystems like gfs2, ubifs, nilfs, ceph,... I'm not sure whether they support blocksize < pagesize at all. Interesting is also NFS which may care but I don't understand its ->page_mkwrite() handler good enough to judge.”

    I’d rather use a file-system that doesn’t need constant fixing.

    As far as popularity goes, Debian’s “popularity contest” data shows 2861 installations and rank 6776 out of 147K packages tracked. For comparison, mariadb-common ranks at 6690, so jfs is definitely in the fight.

  14. oiaohm says:

    –It’s FLOSS. Users can support it one way or another.–
    True but the question is will anyone step up to-do it. Users are free to choose not to support something just as much as they can choose to support it.

    For now using JFS is a question mark.

    Percentage of FOSS stuff become deprecated. JFS seams to be heading down the deprecated path.

    — That means it’s well tested and probably relatively free of bugs.–
    The problem is compared to the likes of XFS that still has on going testing and age behind it has lower number of faults lower than JFS.

    Robert Pogson basically you have to watch JFS to see if a new team takes over. The fact over 12 months has past since the prior team and a new team has not taken over I no longer trust JFS.

  15. oiaohm wrote, “The quality of JFS is the team that has been supporting it. Problem it appears JFS has lost it support team. You will get away with it for a while. But as systems cycle up to newer versions of kernel it condition may deteriorate.”

    It’s FLOSS. Users can support it one way or another. It was great that IBM was behind it for years. That means it’s well tested and probably relatively free of bugs.

  16. oiaohm says:

    –That means no more new bugs…–
    Very bad wishful thinking.

    Robert Pogson Linux kernel changes over time. So a file system driver need a lead developer to deal with any kernel changes that trigger bugs. Up until 2012 JFS had a full time development team at IBM taking care of it.

    The quality of JFS is the team that has been supporting it. Problem it appears JFS has lost it support team. You will get away with it for a while. But as systems cycle up to newer versions of kernel it condition may deteriorate.

  17. oiaohm wrote, “JFS is a file system I put the words historic next to due to the key staff that use to maintain it have basically jumped ship. You will also find IBM has been getting rid of their JFS pages as well. Fairly much they don’t want anyone using JFS any more.”

    Good. That means no more new bugs… 😉 JFS has been incredibly stable for me. I’ve had no issues at all with it, not that ext3 gave me many problems either, but there’s just no reason to have both file-systems on Beast. Migrating to ext4 might have been simple but JFS was not difficult except that the power went out after many hours of running a reformat in place process… Eventually, I did a backup/restore cycle that was much faster and prompted me to clean house.

    I would classify JFS as debugged, not obsolete.

  18. oiaohm says:

    https://en.wikipedia.org/wiki/JFS_%28file_system%29
    JFS has a historic bug in it journaling design is why it failing out of favor.
    Ext3/ext4 and xfs journaling is better designed.
    Dave Kleikamp who use to take care of JFS at ibm now works for Orcale and works on XFS these days.

    –Dave Kleikamp, a member of the IBM Linux Technology Center and JFS Core Team–
    Wikipedia says this
    http://patchwork.xfs.org/patch/6008/ but 2013 patches from Dave Kleikamp are Oracle.

    JFS is a file system I put the words historic next to due to the key staff that use to maintain it have basically jumped ship. You will also find IBM has been getting rid of their JFS pages as well. Fairly much they don’t want anyone using JFS any more.

  19. oiaohm wrote, “So hardware support of debian lenny with a few minor setting is equal to latest testing.”

    In the last few days I had the opportunity to explore Beast’s file-system. I had a mixture of ext3 and jfs, I guess, because I made some mistake back in the day. It’s the original file-system I created when Beast was installed of AMD64 back in the day. I know because of the creation dates of files. I had a devil of a time converting from ext3 to jfs because stuff would not fit… As it turned out I had archives of ancient releases spinning around, even partial repositories that are no longer of any use and databases I haven’t used in years. I even had complete backups of complete backups… Apparently, I had moved stuff into sub-directories so that I could do partial backups. So, as it turned out, I actually cut the usage of my largest partition in half so it would actually fit in another drive on Beast. Originally, I tried to SSH it across to another machine but the power failed shortly before the backup could complete. I actually decided to throw out stuff I created for teaching and haven’t used for years, like my local copy of Wikipedia from 2005. It’s just not worth the space. I may not like contributing to Wikipedia on the web but it sure is more useful than my subset. Then I discovered that the burden of xz compression versus gz just was not worth the time. Presto, Beast has everything JFS except for memory cards. I’m not sure there’s any performance advantage of this, but I’d just rather use a mature file-system than one that keeps evolving like GCC. I guess when I replace Beast’s mobo soon, the old file-system will still work with it. It’s more than a decade old and still going strongly.

  20. oiaohm says:

    The fact you can roll out the new kernel on old debian ahead of doing a full distribution upgrade allows you to locate computer that will have issues while still have the boot option so the user can still access their machine.

    RHEL, Centos and Scientific have the same properly of being able to take the newer release kernel back versions without breaking everything.

  21. oiaohm says:

    https://support.microsoft.com/en-us/kb/2801679
    Have you never heard don’t throw stones while standing in a glass house.

    Windows has had SSL issues. Microsoft gets worse.
    http://www.zdnet.com/article/microsoft-reveals-windows-vulnerable-to-freak-ssl-flaw/

    DrLoser this is the reality. Microsoft is going to try the hide from you the fact their software has flaws. So if you have disabled updates for any reason you will not be informed what security patches you need to install. Debian SSL issues are public and since you are informed you can take mitigation measures.

    Please note Microsoft own research team was the one that found the freak flaw. The fact Microsoft claims that at the time of release they were not aware that Windows was effected. This means 1 of the following events.
    1) They really did not know because Microsoft Security Research team is only a PR department researching the competition to produce bad press. Bad for Windows Security because Windows not being audited properly.
    2) Microsoft intentionally covered up a Security flaw.

    Talking about SSL issues is how to give Microsoft a major black eye.

    –So, your recommendation is that in 2009, Munich should have adopted Debian Lenny?–
    Munich started on Debian DrLoser. They changed to Ubuntu because due to it popularity and store it was thought it would provide more software. Welcome to the sad world of reality. Ubuntu has turned out to be more unstable than Debian. Mostly because Ubuntu does not have the QA resources Debian has. Debian has full servers running 24/7 making sure packages are built properly. Scientific, Centos, SUSE and RHEL distributions also run QA servers like this and they all had this by 2004 so Ubuntu is extremely late tot he party. Anyone choosing Distrobution based on QA process would not have chosen Ubuntu in 2009.

    Debian you can cross kernels between releases. So Lenny can have the latest debian Testing kernel installed if required. Its one of the thing about Debian. A distrobution release version in Debian will be tested that it works with all Debian kernels newer than it that are maintained. Due to Ubuntu custom patching and lack of QA you cannot swap kernels 100 safely between releases. This makes your hardware compatibility support way worse and causes you way more effort with Ubuntu compared to Debian. Yes with Debian you may need to back port Mesa but debian has the backports project that does that. Yes you can still install current Mesa on lenny and not have to build it yourself. So hardware support of debian lenny with a few minor setting is equal to latest testing.

    Most Ubuntu update issues trace to a package modified by Ubuntu Staff. Ubuntu has a QA issue. There need to be a serous focus on getting FOSS projects to use automated QA solutions. There also need to be serous consideration to choosing Linux Distributions with a proper QA process.

    Basically popular does not equal properly performed QA process.

  22. kurkosdr says:

    ” What, precisely, do you mean by “support?” ”

    Psycological support to help vict… users cope?

  23. DrLoser says:

    Because, if you are not running Lenny on the Beast right now, Robert, then that “Five Years Long Term Support” theory of yours is not worth the paper it isn’t written on, is it?
    What, precisely, do you mean by “support?”

  24. DrLoser says:

    Perhaps I could have used the expression, “installed base”, instead.

    Perhaps you could have, Robert.
    But you didn’t, did you?
    And since you did not, I can only take you on your word. Which was clearly preposterous. Try harder next time.
    Of course, Munich could have taken your approach, which is to adopt a bare-bones Debian system and thin clients everywhere. Since they didn’t, you appear to be at odds with your favourite municipality on this one. Perhaps you should write them a strongly-worded letter?

    They would be much better off using Debian who take about 2 years getting the bugs out and releasing LTS good for 5 years.

    So, your recommendation is that in 2009, Munich should have adopted Debian Lenny?
    They would have loved all those SSL issues.
    And, just out of interest, are you running that particular “five year LTS” version of Debian now on the Beast?

  25. oiaohm says:

    Robert Pogson its called miss allocation. Surveys say 1 in 10 machines at most need Windows of course that presume you have at least 10 machines preferably closer to 100 machines.. Problem here is between 2-3 in 10 will most likely be allocated Windows. This job might need X Windows application then it turns out that job does not need X application at all causes the 1-2 extra. Yes 10 to 20 percent variation from ideal for cost. Same thing happens when you have OS X deployed in business for X application reasons not all X application reasons turn out to be true.

    One of the causes of the extra allocation is the Bling factor. People want to say to other staff they have X software even if they don’t ever use it. So they claim software is required for there job when it not.

    Floating between 70-90 percent is kinda normal for a Linux migration based purely on thick client. Many different studies by IBM and others has found this. Strictly managed migrations have been able to get closer to 95 percent Linux.

    Munich 22 departments each department is in charge of over 1000 machines. Its just perfect size of miss allocation to sneak in. Who is going to want to have to audit 24000 machines for correctly installed software for job roles and location.

    One of the problem with thick client solutions is the fact that a User to use a bit of software it has to be installed on the machine in front of them. Munich is a thick client model for everything. Areas that have got to 95+ percent Linux have had thin-client model for a majority of Windows software then only very select machines full Windows. Ok thick-client installed with FOSS is not costing the bottom line so you not worrying about license costs as much.

    Munich is not an example of optimal Linux-Windows deployments. But it is example that even if you don’t do optimal cost saving are there for large enterprises.

    Robert Pogson basically nothing in the Munich numbers are a surprise. City of Largo in one report was 97 percent Linux when it award for being forward looking. But its not the only Linux migration running 9o percent + bracket. All the migrations 90 percent+ have thin-terminal Windows of some form. This is why all the harping on about Munich is not really showing the full story.

    70-90 percent is where you expect a thick client Windows/Linux solution to stall and hover. Thin Client Windows/ Linux thick or thin somewhere 70-99 percent it will stall and hover . If it does not get to 60 percent the project will most likely fail. This is found in survey after survey of Linux migrations. After the Linux migration has cross 60 percent its normally too costly to remove and replace with something not Linux.

  26. DrLoser wrote, “Really, Robert? Something that completes, or makes perfect?
    Sounds like a boffo idea to me. Time to give up on your less than perfect Debian distro and spin up the pure goodness of LiMux!”

    Complement: “Full quantity, number, or amount; a complete set; completeness. [1913 Webster]”

    Perhaps I could have used the expression, “installed base”, instead. It’s interesting that when Limux began, they were working on ~12K PCs but by the time they actually migrated much the target changed to 15 out of 18K, and now it’s 18 out of 24K. I find it puzzling that M$ should be gaining share but then we don’t know for sure what the numbers represent. Perhaps it’s desks or chairs rather than PCs. I think the long term goal is to reduce dependence on M$ as much as possible. Perhaps they rolled out one too many “must-have” applications… All over the world, new systems are going with GNU/Linux. I don’t know why Munich doesn’t follow that pattern. They are doing so much work just to lose ground to M$.

  27. oiaohm wrote, “I have to wonder if most of Linux Desktop problem is not IT personal being incompetent and not able to perform TCO or effort to set stuff up right.”

    In my experience, GNU/Linux succeeds where someone who knows how to do GNU/Linux has the backing of leadership in the organization. Otherwise it’s very difficult to convince a nontechnical leadership that change is worthwhile. Sometimes the bottom line persuades them, but that is rare, especially in government where the power to tax covers M$’s ass. In business, as long as profit can cover the cost of IT, the folks in leadership often don’t care how it works. What is needed is someone at the top that knows there is a better way and makes it happen. Many new rollouts choose GNU/Linux because it makes lots of sense no matter how you look at it. Lots of folks don’t want to change systems they think run properly even if that’s not the case. Where I last worked, some folks thought it was just fine to use XP even if the mean lifetime of a PC was less than a year with it and GNU/Linux just kept running. They think that way because it’s someone else’s problem. The most extreme example was a lady who didn’t want me to “fix” her PC even thought it took 5 minutes to respond to a click. She was more afraid of losing her data to me than to malware… It’s irrational but, strangely, one prerequisite to being in a position of leadership is one way or another to be irrational. Take The Trump for example… Imagine that guy with his finger on “the button”. ~15% of USAians think that would be just fine. Should folks who think like that have “root”?

  28. oiaohm says:

    This is what is so funny you here the stories about issues with Munich

    http://davelargo.blogspot.com.au/

    Largo is a population of less than 78000.
    Has run Linux in thin client configuration for over 8 years.
    City of Largo total staff of 899. Yes Munich IT department alone is 1000 staff with over 33000 staff using computers for something. City of Largo IT team is less than 10.

    City of Largo is Gnome 2 and Munich is KDE as desktop environment.

    The thing here here is City of Largo provides one of the QA servers to the Libreoffice project. They are not big enough to provide a development team but they can provide a server running 24/7 checking for issues with submitted patches.

    Yes the City of Largo is able to achieve the licensing and operational cost savings. Core distribution of Largo operation is Debian with forward ported Gnome 2 because they don’t like Gnome 3.

    That the thing when you start going into each of the Linux deployments world wide they are all different and large percent have been successful.

    I have to wonder if most of Linux Desktop problem is not IT personal being incompetent and not able to perform TCO or effort to set stuff up right.

  29. dougman wrote, “Two people whine about NOT having ADMIN rights”.

    Yep. The easiest way to shred an installation of That Other OS is to give admin privileges to folks who just want to do their own thing while expecting others to “fix” things. I’ve seen too many animated cursors and “auto-run” malware and random executables installed in the name of setting things up the way the user wanted. Then there are the folks who like to click on attachments and to say nothing of installing unlicensed software…

  30. Mac Taylor wrote, “This misdirection occurs every couple of years to draw attention away from MS failures”

    I find it interesting that one of the writers earlier supported the move to GNU/Linux and that now both issued these letters and conveniently went on vacation, like tossing in a hand-grenade and running away… Purpose seems to be to do damage rather than to improve the situation. Who are these elected officials serving?

  31. dougman says:

    Two people whine about NOT having ADMIN rights….BOOHOO.

    I’ve run into that mess before, and have to explain to C-level types that I would no longer be responsible for that device.

  32. Mac Taylor says:

    On the surface, it’s just 2 people who really want admin rights to their government issued laptops in order to facilitate Skype and MS Office installation; not going to happen.

    This is the link to the original pdf signed by the requestors;
    http://www.ris-muenchen.de/RII/RII/DOK/ANTRAG/3745830.pdf

    Sabine Pfeiler, Stadtrat
    http://www.ris-muenchen.de/RII/RII/ris_mitglieder_detail.jsp?risid=3312549&periodeid=3184778
    Otto Seidl, Stadtrat
    http://www.ris-muenchen.de/RII/RII/ris_mitglieder_detail.jsp?risid=1418516&periodeid=3184778

    This misdirection occurs every couple of years to draw attention away from MS failures.
    In this case Windows 10 (the toaster doesn’t fly so well now).
    http://www.ocsmag.com/2015/08/24/no-munich-is-not-considering-ditching-linux-and-going-back-to-windows/

  33. oiaohm says:

    https://wiki.debian.org/LTS
    DrLoser like it or not Debian processes are changing with Munich and others feedback and support. 2 year release cycles too short so now 5 support time frame.

    https://wiki.debian.org/LTS/Team
    This page is interesting. Everyone with () with company name is paid to be there. NIH syndrome is unfortunately common in companies and governments.

    The base of Munich is a +80 percent license saving. That for something the size of Munich operation that covers the Linux support staff with left overs.

    Agent_Smith things change over time Systemd is becoming less and less of a problem. Of course its not perfect but sysvinit/upstart before systemd was not without is share of issues.

    Yes all the hope the anti-Linux trolls get about Munich converting back shows they don’t understand the problem because its not going to happen. Financially it cannot happen. This is the problem for Microsoft once a Linux migration crosses the 70 percent point conversion back never seams to happen no matter what Microsoft offers.

    The Linux issue is part a developer numbers issue. The more companies and governments that convert the larger the developer team to improve and maintain Linux becomes.

  34. DrLoser says:

    Oh yes, and “agitators.” Lovely little reminiscent phrase, that.
    What next, “Cosmopolitans?”

  35. DrLoser says:

    The complement of PCs has risen to 24K

    Really, Robert? Something that completes, or makes perfect?
    Sounds like a boffo idea to me. Time to give up on your less than perfect Debian distro and spin up the pure goodness of LiMux!
    Except that you won’t, will you?
    So why expect an entire German city to deal with the sort of crap that you would never touch in a million years?

  36. ssorbom says:

    Munich isn’t just running Ubuntu. They are running a customized version of Ubuntu. The Linux Action Show guys were groaning about this because their custom distro looks ten years out of date and gives Linux a bad name for people in Munich. I agree Debian might be a good choice here. That or CentOS, or RHEL.

  37. Mats Hagglund says:

    It’s well known fact that people don’t purchase operation systems, they just buy computers. As some studies have proved many people are thinking that Google, Dell, Internet Exploarer or Firefox are operation systems. If governments forced folks use some Linux operations systems there won’t be any kind of revolts on streets. Especially if they got their computer with OS+applications 20% cheaper that with Wintel.

  38. Agent_Smith says:

    They should stick with Slack, that is. Debian ? Seriously Robert ? That Debian infected by systemd (which makes it almost unusable, as you noted in previous posts).

Leave a Reply