The Document Foundation on City of Munich’s discussion on returning to Microsoft | Hacker News Reader

The Document Foundation on City of Munich’s discussion on returning to Microsoft

blog.documentfoundation.org

136 pointsmksaunders66 comments

18 to 28% is not a minor percentage of users. I'm a fan of LibreOffice, and was an extremely early adopter and contributor to the project, being one of the people that pushed for and created the community council at the time. Brushing these kind of numbers away isn't going to help, and they will be hit hard on that point by the opposition. The better approach will be to adress the mismatch head on, with an actionable plan on how to solve these issues.

I worked hard, in a professional capacity, for the adoption of LO and Desktop Linux in various government departments around the EU, and the agression of commercial players against this kind of push is phenomenal. Eventually I moved on professionally but the scars are still with me. One powerful message this whole thing left me with is that you cannot do this kind of thing, and hope to have a career in professional services. I was forced to drop out of public participation with OSS projects, and never really went back.

>> the agression of commercial players against this kind of push is phenomenal

Theses are extremely strong statements. Could you elaborate (if it doesn't put your career in danger, of course) ?

I'm no longer playing that game, so sure.

I was contributing for some time to OpenOffice, and at one point starting working for a systems integrator, specialising in Linux and Open Source solutions. At this time, OSS was just coming out of academia, and enterprise had no interest in all this hippy bullshit. They needed a techie in a suit to come and tell them the how, why, and how much OSS would help and cost. As soon as a client would mention anything related to Linux or OSS to the salesdrone, I would be dragged into the discussion. This was a pretty cool job: lots of very large clients and projects which suited my background (designing and implementing very large scale infrastructure) and interests (Linux, Open Source). I was an active and vocal contributor within the OpenOffice project, recognising it as one of the key enablers for a possible Linux Desktop. Sun sent me increasingly strongly worded requests to "tone things down" - I was pushing hard for the idea of a community council, and Sun didn't like the idea of giving up control. One of the results of this was that I set up OOoForums.org, since we didn't really have a space to freely interact with end-users on the "official" site.

One not-so-fine-day I'm called into HR. VP's present. Heavy meeting. No warning. My employer had a strong and profitable relationship/partnership with Sun, we were selling big Sun iron into enterprise to the tune of millions per year, and Sun complained about my behaviour to my employer. I was told in no uncertain terms that if I wanted to stay happy at my job, I would do as I was told by Sun via backchannels. If not, I'd lose my job. Having a family to take care of, I dropped out of the whole scene the same day. (this is now about 12 or 13 years ago)

I left this company not long after, and did a project to deploy Linux Desktops with OpenOffice not long after. A Government dept, about 100 seats if I remember correctly. New building, new pc's for everyone, and we sold it as "A new version of windows". We spent a long time making KDE look like windows, and made sure everything worked. It was awesome, it cost a fraction, and the users loved it. I was hired by a sympathetic client who specifically wanted this. There were upper management changes during the project, and the new head let me know that my services were no longer required, since he was flown to Seattle by MS (from Europe) and they "explained" to him that MS was the way to go. I pointed to my (extremely solid) contract, and let him know he could do what he wanted after delivery. I was subjected to significant pressures for the duration of the project, and it really wasn't a lot of fun. A week after operational delivery, IBM and MS waltzed in, blitzed everything, and that was that.

Frustrating business....

My employer had a strong and profitable relationship/partnership with Sun, we were selling big Sun iron into enterprise to the tune of millions per year,

Wonder if your ex boss ever thinks about this now? I suspect there is less demand for Sun or Oracle big irons now.

They are a very large EU-wide integrator, and have partnerships with pretty much all the boxbuilders. The whole saga with Sun hardware was likely just another profit opportunity.

Thank you for posting such a long explanation. It's very valuable to both realize the power of open source (by indirect means of the fear it created) and how hard it is to sell it when it has to fight against an up hill opponent... I am not surprised by these facts but I'm quite surprised by the power of the forces at play. Thanks again !

My pleasure :) The power of OSS is amazing, as is the power of those who believe they stand to lose.

Although they word it poorly, I think the point is that those numbers should be compared to similar software issues experienced on Windows platforms, in which case the benefit of switching back is a lot slimmer. As noted a bit later:

> Incidentally, 15% of users acknowledged severe issues related to MS Office.

I assume these 15% complained about interoperatability with 3rd parties that use the MS Office DOC format.

You have a point in that it's very easy to interpret these statistics in totally contradictory ways, but without further details we are just speculating.

The Microsoft fanboy in my office seems to hate Word.

I just dislike it strongly and prefer to use LibreOffice for cleaning up styling done by others is top of my mind.

" I'm a fan of LibreOffice"

I am not. But you can run MS Office with Wine, if need be. I do this on occasion since it is the only reliable way to make sure the document will look the same at your client.

For all other stuff, I use Softmaker for Linux. http://www.softmaker.com/en/softmaker-office-linux

I had a look at the linked report https://www.ris-muenchen.de/RII/RII/DOK/SITZUNGSVORLAGE/4277...

Apparently, the actual recommendation was to first replace the 4.163 Windows clients in use (77 % Windows 7, 9 % Windows XP/Vista, 14 % Windows 2000 [!!!]) with a single improved configuration, likewise for the different versions of their LiMux distribution.

Afterwards, each department should choose between the Windows and Linux clients based on a cost-benefit calculation. Only in case the Windows client ends up dominating, a reevaluation of the economic viability of the duplicated infrastructure is recommended.

Additionally, the only mentioned software problems from their user survey were caused by varying browser versions (presumably Firefox), their mail program (presumably Thunderbird) not being available on mobile and frequently crashing, the same complaints about their calendar software (no idea what they are using) in addition to lacking copy-paste functionality.

Unless LibreOffice has a calendar component, I don't think anyone complained about it? At least it isn't in the report.

That seems a rather narrow way for Munich to operate. The world has changed since they made their decision a decade or so ago. Microsoft now offers on-premise and cloud versions of all its main programs, with companion apps that also run on Android and iOS smartphones and tablets.

Windows 10 machines now include cloud integration, apps and touch screens, with hardware form factors ranging from small tablets through 2-in-1s and laptops to the Surface Studio.

If they are still comparing one 20th century desktop OS with another 20th century desktop OS, they are missing the point.

I think the article misses a great point. That by using open source they can spend on IT locally! Look at the Extremadura (region in Spain) case for as a reference; they went with open source to stop all spending on IT outside of their own borders and thereby significantly boosted local business.

To me this is a clear case of how lobbying can destroy the value governments provide to their citizens.

They did mention that point:

> Last, but not least, most expenditures related to the purchase of Microsoft licenses will contribute to the GDP of Ireland (where all Microsoft products sold in Europe are sourced from) rather than to local enterprises who support the open source solutions deployed today.

IANAL, but isn't encouraging the City of Munich to source from "local" enterprises vs Ireland a violation of EU free-trade laws? By the same token Ireland could then discourage/block "imports" from Munich and we'd be back to the internecine trade conflicts the EU was supposed to stop.

Free market doesn't mean you have to go with the cheapest solution. It means you have choice. Preferring local firms is ok because they're competing in a free market; they aren't being given contracts because they're local, it's just local firms will be opening up to supply the demand.

For a local government a preference of local business is not only ok. It should be the first choice, the second choice and also the third one. In this case, as there is totally different tax flow involved as the article nicely explains, local business is the only responsible choice for an elected government, IMHO.

Oops. Missed this. Thanks.

Microsoft Germany is located in Munich. Or, well, it will be moved there soon.

I'm disappointed in Munichs' decision-makers in this case. Introducing linux was a great step towards a more open, secure and cost-effective platform to organize the city. The reasoning behind switching back to Microsoft is vague to me, the main factors seem general it-problems not related to linux, users wanting to open their .exe and windows programms (wouldn't be allowed with a Microsoft system as well due to security issues) and the fact Microsoft built a new shiny office.

I think it is because that Microsoft just opened an office in Munich...

I think it may have something to do with whatever makes almost no one choose Linux for desktops.

That 'almost no one' is growing steadily.

https://linux.slashdot.org/story/16/07/02/184229/linux-grabs...

Yes, perhaps one day the 'steady growth' will hit 3%. Perhaps then we can declare the "Year of the Linux desktop".

Here's mschuster91, who worked there, take on the issue from a couple of days ago when it first was posted to hackernews:

------------------

> Does anyone have some examples of the practical problems they experience using Linux, which can't easily be solved in any other way than going to Windows?

I worked there. The biggest problems back a few years ago:

1) So-called "Fachverfahren", basically software for stuff like managing drivers' licenses or other bureaucratic procedures, is written and supplied for Windows only. Often enough that meant Munich had to pay a boatload of money for a Linux port. Or the software HAD a Linux port, but for RHEL or other "enterprise linux" crap distros - which meant using stuff like "alien" or, worse, manual repackaging to make them compatible with Ubuntu. Yay for version hell - statically linked programs were an exception.

2) DRIVERS. There's a lot of custom hardware - special printers for printing on documents like ID cards, fingerprint readers for the new national ID cards, RFID readers with support for said ID cards... you won't believe how much stuff there is. And all of this needs to have Linux drivers and tooling.

3) Employee training. Back in my days KDE was used (and I believe it still is), but it's different enough from Windows that people need training. And there's 35k of employees, most of which don't have any IT experience outside from their Windows computer at home.

4) The computers themselves. Let's just say that the computers in any public agency are almost always horribly outdated. Many users complain(ed) about the speed of LiMux, which mostly was caused by old or underpowered (esp. RAM, given that OpenOffice and Firefox are really really memory hungry) systems.

5) Networking. Depends on the building and agency, of course, but e.g. my school was connected via a 16 mbit uplink, over which the entire Internet traffic went...

I believe the biggest problem LiMux had and still has is a lack of proper funding - especially for hardware.

--------------------

tl;dr: it's more than what you suggested and some very valid reasons. I know some people involved in that decision and let's say it like that: LiMux does not come for free as well. There are plenty of issues that require external consultants for 1500 Euros per day/person which is apparently always an unexpected and unbudgeted expense.

I haven't followed this story but if they deployed an Ubuntu/debian based distro over CentOS/RHEL then they were asking for trouble. In my experience, the commercial vendors who release software for Linux almost always target RedHat-based distros first, and often only.

And RHEL is also not free. So where do they save money compared to Windows? In that case they will be supporting FOSS but not due to it being good "for them" but out of a moral high ground.

Yeah, RHEL is one area I wouldn't spend the money on desktop deployments - CentOS is a much better choice here (it's $300/year for a RHEL Workstation license with support, you can get self-support desktop builds for $50/year but why not just use CentOS at that point).

35,000 office workers for 1 city?

I realize Munich is a pretty big city, but holy crap, that sounds extremely bloated. I can definitely see why they tried to save money on Windows licenses, but perhaps their choice of OS isn't the problem that needed solving in the first place. Maybe their efforts (and money) were better spent on automating many processes (e.g. like you said - printing ID cards should be 100% automated and not require human intervention in most cases)

Or maybe it's total employees, and not only people who sit in the office?

If I were mayor, I'd seriously look at the bureaucratic bloat required to run the city and trying to make it more efficient (Windows to Linux you just shift money around from licensing to retraining and overhead)

This includes all workers, including childcare, public schools, museums and garbage disposal. The administration itself has around 8000 employees. They do practically all everyday tasks: issue parking fines, marry people, register residents, issue residence permits for foreigners, organize all elections, grant building permits, organize the city's own construction, plan traffic, register cars, and so on. Maybe you can streamline a process here or there but there is not much you can save.

Regular ID cards are printed by a central contractor in Berlin (presumably automated). However, emergency ID cards or passports are printed directly in the office with special ink. There is not much you can automate here – it doesn't happen too often but can't be batched as people need them immediately. Just with ID cards I can think of a lot of other specialized hardware: optical scanners for the ID card itself, optical scanner for paper documents, fingerprint readers, contactless readers, automated forgery checkers, and probably more. All of that needs to be supported.

I can't think of many processes that still aren't automated that could be. The only thing is that most internal processes where many employees need to work together are still paper based and a file is often transported physically to a couple of people – for example to have a medium-sized contract at least the case worker, their supervisor, the tender office, a lawyer, and the treasury need to have it on their desk. But paper is going to be replaced with an electronic file system soon (but again – that is probably not developed for Linux and needs to be ported).

most socialist governments have public-sector employment rates approaching 50% (or sometimes more). In my canadian province, 49% of the population works for the government.

That's ridiculous, Munich has around 850,000 employees, so 35,000 is nowhere near your 50%. Do you actually know what socialism is?

Nowhere in my post did I say that Munich was a socialist government.

>3) Employee training. Back in my days KDE was used (and I believe it still is), but it's different enough from Windows that people need training. And there's 35k of employees, most of which don't have any IT experience outside from their Windows computer at home.

I've been seeing this pushed as a Linux disadvantage for at least 18 years now.

So if employees can't figure out how to switch from WinXP or Win7 to KDE, please explain exactly how they're going to switch to Win8/8.1/10 without any training.

If no one is getting training for Win10, then no one needs training for any Linux DE.

I believe focusing on potential technical problems is the wrong approach.

What you have to look at is lock-in or no lock-in. Any responsible IT Manager should avoid lock-ins / silos as far as possible. This then allows to standardise and use multiple suppliers and much lower TCO.

Without having looked into details (but with some knowledge of the Munich IT environment), I'm sure that most of these technical issues stem from the previous lock-in situation with MS.

There are other solutions without migrating the desktops back to MS / upgrading to MS Win 10 - you could for example virtualise the apps with issues (Standard approach & working very well in large environments). But then Munich could not be locked back into the MS Stacks so easily.

And MS and Oracle are very good with locking their customers in and collecting vast amounts for that.

BTW - I've seen large government environments in the UK with completely virtualised application delivery / desktops to enable staff to work from home, BYOD, hot-desking etc while at the same time securing the organisation's data e.a At that point which desktop OS you are using has only limited importance.

I work with Linux professionally and generally like it as a server solution, and also as a developer, but if I would be a city mayor, I would not switch to it from Windows in the first place, unless at least the country government would do such a step first.

> In fact, although the proposal associates MS Office document formats with the “industry standard” concept, it should be clear that all MS Office documents are proprietary and obfuscated, and therefore inappropriate for interoperability.

Sorry, but this is utter BS. MS Office is a de facto standard, and LO is not compatible with it - it can't correctly open more complex MS Office documents. I should know, since I periodically evaluate it. So the reality is, that MS Office ensures much greater interoperability as LO. Or maybe I am just missing something and we are talking only about server solutions? For example - mayor secretary receives a document from some government agency, in docx format, and LO can't correctly open it. Now what is the fallback solution?

You are dumping problems related to MS (well known) monopolistic behaviour. Office format has been notoriously hard to replicate (on purpose). That is not a failing in Open Source, it is a failing of Microsoft.

I realize that businesses must run daily operations and that requires them to bite the bullet and pay the "Microsoft tax". A municipality should think a bit about if some operational problems have a price tag of 90 million euros annually.

Yes, you are right, and it is good that they though about that and tried the alternative. But I still think the issue of LO not able to read MS documents is a real problem.

Like for example my sister. Her husband works for Redhat so he enthusiastically installed Linux on her notebook, LO and everything. She worked as a professional writer, and it was a real struggle... some documents she received wasn't readable, then they couldn't read some document she sent them... then spell-checker wasn't as good as in MS Word (non-English language), etc. The cost of this solution was quite high, and she eventually gave up and installed Windows.

I am not trying to trash Linux or LO, it would be great to have open-source alternatives to everything, but in certain areas (Office, Adobe products, etc), such alternative doesn't exists and fighting against reality doesn't make it.

> some documents she received wasn't readable, then they couldn't read some document she sent them...

I like that! I send you a document you can't open: your fault. You send me a document I can't open: your fault too.

That's inevitable when you're going against standard practice in the industry, and refusing to fit in with everybody else.

Try the same thing in any other industry and you'll get the same result. "You created the problem: you deal with it."

> some documents she received wasn't readable, then they couldn't read some document she sent them

Which is survivable if your time has no value. If you're paying people serious wages to do this stuff, the costs are dramatically higher than just buying Windows and Office.

When I was in the full-time editing business, sending me unreadable documents more than once would not be a good way to get further commissions.

Specifically, someone on $50,000 a year is earning roughly $25 per hour. Putting them through the painful time-wasting problems of dangerously semi-compatible software is a huge waste of money.

Office 365 easily pays for itself in saving a few minutes per day.

> (on purpose)

Any evidence for this? I specifically remember one article talking about Excel's file format that was engineered in a way as to run on extremely resource-limited computers - those available when office was first released. Any complexities had to be built on top of that legacy base, and be backwards compatible.

That seems to more than explain the file-format complexity, without unsubstantiated allegations of deliberate obfuscation.

> LO is not compatible with it - it can't correctly open more complex MS Office documents. I should know, since I periodically evaluate it.

When was the last time? Got much better with the last release.

I tried to fill out a job application form last month. Got the latest OO and Libreoffice, the document was completely messed up - repeated sections, some parts out of order.

Filed a bug, then bought a 365 subscription.

It was a .doc, but this is still used in real life.

The LO team would love to get a copy of that document, if possible.

> Filed a bug

They do.

> mayor secretary receives a document from some government agency, in docx format, and LO can't correctly open it.

Yeah that never happens when upgrading Microsoft Office..

So you're suggesting that a city government should continue to support and basically aid and abet the monopoly that Microsoft has in wordprocessing and spreadsheets? That sounds like you don't like the free market.

>MS Office is a de facto standard, and LO is not compatible with it

Neither is a different installation of MS Office.

Oddly enough, I regularly use three versions of Office -- 2007, 2010 and 2016 -- with zero problems.

> For example - mayor secretary receives a document from some government agency, in docx format, and LO can't correctly open it. Now what is the fallback solution?

Yeah, what do you think could be the fallback solution if one government agency sends a document to another government agency in a secret language, telling them that they have to pay the only company that knows that secret language (and that they paid to translate it into that language) to obtain a translation in a language they can read? Any ideas what possibly could be the solution in this scenario?

The thing about moves like this is they are looked in far too short-term of a mindset. The technologies needed to do $things are an investment over time, one that I have seen repeatedly backfire in choosing proprietary software the longer that time goes on. It's not just about the budget for the next year or five, it's about freeing yourself from the technical limitations, the legal limitations, and by what I call death by a thousand cuts licensing fees.

Sure, the initial pain of a relatively rough transition such as this is going to be higher than expected when this isn't taken into consideration, because you are going to have to find or pay to dev all those tiny little proprietary programs that have attached themselves like a virus to your budget and replacement them. If anything, I would say the pains of Munich make an even stronger case for breaking through the spell of proprietary programs.

I do have to say though, I think Ubuntu was a bad choice for the distro. That being said, as I said in the other related thread, and as seems to be verified by the poster who worked there, I highly suspect half of the problem is a lack of budget and an overworked/understaffed IT team.

You know, I just thought of something, if "The user controls the program, or the program controls the user." Could you say then, that "The government controls the software, or the software controls the government."?

I get the impression that this has become a party-political issue, and the change of heart is because the other party is now in charge, rather than the original decision-makers finding that that the switch turned out worse than they expected.

As much as I'd like to see more support and success stories for Linux/LibreOffice in public administration, as I think it's the right thing to do, putting statements like this into the document makes me question their motivations:

> In addition, according to estimates provided by Green Party councillors, another 15 million euros should be spent to replace or upgrade PCs which are perfect for a small footprint operating system such as Linux, but cannot support even a Windows 10 basic configuration.

If they compare that really with KDE4 on Ubuntu 12.04 I'm not sure how they come to that conclusion. Windows 7 or 10 is just about as fast or for some tasks even faster than a Linux Desktop with a full blown DE. In my experience feels Windows even faster on older machines due to less lag (I might be bullshitting myself here).

And without enough RAM Linux is a pain to use - as is Windows.

Lately I'm asking myself if that Linux advocacy myths that are pretty widespread hurting more than they help.

Yes Linux is fine, it's usable, it's fun. But in my experience using Windows/Linux with DE (Gnome, KDE) on low end hardware (P4, Core) Windows has an edge and feels faster.

Or they are running a minimal i3 setup but I doubt that.

I've dealt with the ICT department of Munich many times and was always impressed with their expertise and can-do attitude. This feeling of power and independence is more valuable than any of the monetary calculations.

Free software is about empowerment.

The functionality required of office suites in government is usually simple. A simple office suite with dedicated people can create a happy working environment.

I'm a daily witness of the drudgerous mindset that is the result of working with a locked down computer. It saps the will to live and makes people irritable and hostile towards the citizens they are meant to serve.

Choosing freedom is the best option.

If Munichs public service personal has issues with Linux and Libre Office, they need a proper training, not a switch of platforms.

Without training nobody in this world would be able to properly use Microsofts products as well. And this is true for every new version, as the comments correctly mention.

How will training help them if their business applications are not available for Linux?

Well, I do expect the availability of applications has been checked before moving to Linux. If not, ... bad management (not impossible that the move was planned badly though).

But if you know exactly what you need, you can also develop software or extend functionality of available software. That's one of the advantages of the Unix processing philosophy and open source software.

Missing from this letter is a clear statement about what makes their product better.

This is the only libreoffice marketing any of us will read this year (and this forum has a lot of LO users, I suspect -- me included). Talk up your product!

This statement seems very defensive to me. I would have liked to see some kind of discussion in trying to learn from Munich's experience to make LibreOffice better.

It's Open Source... you don't need spend time convincing people your product is better, as you don't get revenue from users. What OS projects need is community and participants, which you gain by being open to criticism and game for new efforts to improve.

Reply on news.ycombinator.com