The data presented in the article provides compelling evidence that the answer is yes -- i.e., the migration's recurring savings exceed its upfront costs:
* The city no longer has to pay for license upgrades, thereby eliminating a major recurring cost forever -- savings of nearly 3 million pounds every three to four years, according to the article. That's huge.
* The city no longer has to upgrade desktop software or hardware as frequently, reducing another recurring cost forever -- also huge.
* Most surprisingly, the city claims its IT department is fielding considerably fewer user complaints with Linux than with Windows, reducing another major cost forever. Labor-intensive IT support is always the costliest component of operating a corporate desktop, so that's also huge.
If the recurring cost savings claimed by the city of Munich are accurate, every budget-strapped city in the planet should be seriously considering this kind of migration to desktop Linux. It makes a lot of sense purely from a financial standpoint.
[UPDATE: I toned down the language and corrected key figures, which were off by an order of magnitude due to an incorrect reading of the article. I also materially changed the bullet point regarding support costs, as the article itself was slightly misleading on the matter. THANK YOU Xylakant and luser001 for pointing out my errors!]
Most offices these days do use SharePoint heavily and some of the BAs started their hands on InfoPath to make forms for data entry to SharePoint.
To make matter worse (or sweeter, depending your POV), these whole thing can be acquired cheaper by using Office 365 (their cloud version of Office) plus certain Office 365 offerings automatically gives you Office 2010 Professional Plus for free. Sweet deal if you ask me.
Some of us would like to see Linux to win the battle especially when it comes to cost perspective. Unfortunately, Microsoft took notes and they had adjusted their licensing cost as well.
For example: in the K-12 sector, I heard schools are getting 90% discounts. Ninety percent discounts.....
You must not forget to factor in the huge pain of living with SharePoint. A good friend of mine, heavy Microsoft user (and advocate) regularly complains about downtime and data loss because of their SharePoint install. They were forced into it because Microsoft is one of their most important clients and demanded that they use it (or they'd move their business to a competitor)
Where we did have big problems was finding developers to work with it; we actually found a few that had worked with it but deliberately avoiding listing it as a skill because they did not want to ever work with it again. (in Sydney a few years ago)
I can't comment on that particular issue.
You're absolutely right about Microsoft Sharepoint (and InfoPath): it makes switching away from Windows much more painful and costly. FWIW, I have no doubt there are a lot of smart people at Microsoft thinking 24 hours a day about how to increase switching barriers for customers.
As to Microsoft's licensing discounts in key markets, they're not new. Ballmer traveled personally to meet with Munich's mayor and reportedly offered similarly crazy discounts to prevent them from switching to Linux. (See http://www.infoworld.com/t/platforms/microsofts-ballmer-figh... and http://translate.google.com/translate?u=http://www.heise.de/... for some details.)
Regardless, the important question (for budget-strapped cities and governments) remains the same: can they save a lot of money by switching to desktop Linux? If the savings claimed by Munich are accurate, the answer is a resounding yes.
That makes plenty of a sense. If you force people to use Microsoft products while they're young they are much less liable to switch to another platform while they're older. They're use to their tools and people are lazy. So heavily subsidizing educational uses of software makes plenty of sense.
God forbid we ever lived in a hacker culture where kids in K-12 were forced to learn how to code and actually understand how computers worked as part of the curriculum, then Microsoft would be in real trouble.
Would forcing them to learn how to drive be a sure way to make them hate driving?
If I recall correctly, according to the director of IT of a large, local non-profit, they get access to the full suite of enterprise Microsoft products for something on the range of $1000 per year.
At that price point, saying 'open source can save you money' is meaningless, because they have already made investments in Microsoft IT expertise.
- The recurring costs for licence upgrades are said to be 2.8 Mio Euros every 3 - 4 years. However, there are other cost factors that need to be taken into account: migration and training costs for the new windows and office versions and cost that are now saved since they took the opportunity to unify some of their document management. That's not 10 mio pounds per year, but still a sizable amount.
- In the original document, no amount is given for the cost savings for the hardware that's not replaced.
- There are no statistics given for the number of support queries in terms of linux vs. windows. There is anecdotical evidence but they response explicitly states that full and valid statistics will only be available in two years. The number cited is the total number of support queries for the linux workstations over time. It is cited as proof that the pain points from the beginning of the migrations have been resolved and that the number of support queries is dropping sharply even though the number of linux workstations is rising. The number is also said to include all kinds of errors, including network, server and database errors.
Still, the cost savings are totalled at somewhere around 3.8 Mio Euro for the whole period and since most of the up-front investments have been done now the number can be expected to rise. It's not as huge as you make it sound, but it's still big. And most importantly its a proof that this can be done without disrupting everything.
(most numbers taken from the original statement at http://www.muenchen.de/rathaus/dms/Home/Stadtinfos/Presse-Se...)
1) Navigate the licenses
2) Talk to the 3rd-party re-sellers (their response time varies)
3) Approval to purchase
These are "hiccups" that will slow down all projects were you had to go through them (and it is uncommon situation in Microsoft shops).
I agree with your other points.
The maximum number of complaints was 70 per month
before the beginning of the switch to LiMux. After the
number of LiMux workplaces increased from 1,500 to
9,500, the maximum number of complaints per month
dropped to 46.
Uh, what? Complaints can change at any time and due even to "small" changes. Also, there could be changes in the opportunity cost of running Linux vs other operating systems in the future which could be realised/visible to ordinary users and thus cause a rise in complaints. But yes, for now, if the facts are right, they have saved themselves significant support cost.
Rest of your summary deductions are sound.
When I hear that metric I think of a homegrown app that was supposed to upload files from PeeCees to the company's mainframe. At the annual IT meeting ...
"And since implementing change foo and bar in the system user complaints to the special helpdesk have declined 90% .."
My boss stood up  "That's because the users got tired of calling and not having anything fixed. It's still broken."
The manager/presenter shot daggers at Eric . Cross-hall verbal sniping started. Entertaining.
 He was retiring in a few months.
It starts with 'Bobby in accounting told me to X when Y happens' and snowballs from there.
It's horrible on productivity and when IT finally does try to fix it, it's often much, much worse. (Though, when the user gets it right, IT doesn't have to fix it after all.)
It's also horrible on morale.
Exactly what happened to me today.
An engineer  opened up a ticket because some java code he was developing to insert attachments to our BOM system was returning funny errors.
Which is fine except he'd somehow gotten credentials for my production system and was testing his code against live bits.
Hooray for initiative, but good lord, fella: how would you feel if I ran out on the floor and started testing new widgets on equipment we're building for customers?
 Not software but electrical.
Take 12000 civil servants losing one hour a week on fighting the tools, and you get to 11.7 million pretty fast.
At least with Outlook I solved the problem backing up via IMAP to my own server, who kept the mailbox in a very easy to manage maildir folder. The usual Office workaround - opening with OpenOffice and saving back to Office format - never worked for me after 2008 and backups had to be used.
It's not that you could not get it to work, just there was a vary small 'happy path' and doing just about anything else caused something to crash. EX: Open Excel, if Excel is not opened PowerPoint will crash when you try and edit one of these slides.
PS: I suggested just using images and attaching the originals, but the client wanted to be able to resize things / change the graphs titles etc.
As with Outlook's PSTs, there were no apparent cause for the corruption - it just happened.
As for the other projects and teams, we have significantly reduced the use of Project as we moved our tracking to Jira and use Confluence to keep almost all project artifacts. Despite the fact we still use Exchange for e-mail, I dropped desktop software and moved completely to OWA, with Evolution doing backups over IMAP into a local maildir rsynced to a durable datastore.
Frankly, I would go so far as to say that a Linux desktop OS is really only appropriate for enthusiasts or those with minimal needs (e.g. email). For just about anyone else, there is simply too much potential hassle involved in getting the basics to work (external displays, anyone?).
You are right with Excel although. However, much of the "power" of Excel is really just needed because people are overusing it. Stop making "databases" out of Excel sheets and implement good data analytics system on the server and you'll see that libreoffice is more than enough.
The reason I feel Linux to be appropriate for novices with bare-minimal needs, is that when operating within very narrow constraints the most popular Linux distros are both efficient and safe. For basic functionality, the price can't be beat.
However, hassles really do begin to manifest themselves once users begin to really engage their PC in any number of ways. The vast majority of PC users fall into this category. In other words, they have non-minimal needs but are neither enthusiasts nor developers or otherwise employed in IT full time.
WiFi, external displays, software limitations (e.g. video editing) are only a small number of issues that can and will manifest themselves. For the vast majority of PC users, this constitutes an unacceptable hassle.
Enthusiasts, on the other hand, know what they need and how to fix the above if necessary. I personally think OSX is a superior choice. I value productivity, and don't consider (for instance) mucking around in xorgs to be a good use of my time. However, I respect and understand those who choose *Nix over OSX.
edit : Can one of those who felt this was worthy of downvotes explain what they felt was factually inaccurate?
> The vast majority of PC users fall into this category.
This comes up all the time on HN. "The vast majority of users think X", "the vast majority of users live in Outlook", "the vast majority of users only need an iPad", etc. etc. The reality is that you have no idea what the vast "majority of users" want or need, you are projecting your own biases and micro-observations onto the world.
Actually I think we need to get over the idea that there is some 80+% of users who all need/want the exact same thing. Seems to me that there are probably a lot more categories out there than the black/white you're either a total novice or a power user world that is constantly constructed here.
A number of these issues that you mention are only present with a lack of planning. If you go out and buy a computer with the intention of running Linux on it, you plan and you pick a computer with good compatibility. With each release of popular distros such as Ubuntu, compatibility improves massively. My Dad's laptop used to have issues with WiFi a few years back, now it connects fine (faster than Windows too) in Ubuntu 10.10.
External displays? I've never had an issue with this and I'm actually typing this on a dual-monitor Linux setup. If you have an Nvidia graphics card then this should be fine. Again though, if you plan this isn't an issue.
Software limitations, granted. Just like for OS X, there are professional wine layers which offer Windows software on Linux. But sure, that's not perfect. There is quite a bit of good quality software (plus, obviously, web applications work fine) and especially for development work. But this is definitely an area that needs improving.
Personally I'm waiting on Canonical to release their own computer (or partner up with a hardware provider) for then there will be no issues with hardware compatibility or many of the problems you mention, for that is the reason they exist: attempting to run Linux on a device that was not designed specifically for it.
Just my $0.02 anyway.
Couldn't have put it better. Those who complain about how linux doesn't "just work" often overlook this point.
Personally though, I don't think Canonical coming up with their own computer isn't really going to help; I wish that there were a hardware vendor that designed really beautiful computers so that I could run linux on it (Currently, the best design is by Apple)
> However, hassles really do begin to manifest themselves once users begin to really engage their PC in any number of ways. The vast majority of PC users fall into this category. In other words, they have non-minimal needs but are neither enthusiasts nor developers or otherwise employed in IT full time.
The vast majority of PC users do not fall into your category. They will be people who need to read email, do a little bit of simple web-browsing, create and open simple office documents (the vast majority of which would be fine with Libre-Office) and maybe run some awful accounts / stock / ordering software.
> WiFi, external displays, software limitations (e.g. video editing) are only a small number of issues that can and will manifest themselves. For the vast majority of PC users, this constitutes an unacceptable hassle.
WIFI is fixed in Linux and has been for a while, you don't want to run wifi in a business, and if you are using it anyway you'd have it set up by IT.
I'm not sure what you mean by external displays. If you mean dual monitors than very few people need two monitors, and those that do would have them set up by IT, and they'd work with whatever software being used. If you mean projectors then I have no idea what the Linux situation is. I accept it might be a problem, but that's a problem faced by a very small number of people.
A tiny number of people need to edit video. In the context of a local authority we can say that almost no one will be editing video.
OS X, in the context of local governments trying to save money, is - and I say this as politely as I can - a stupid suggestion. The hardware is much more expensive, is a lot more desirable to thieves, it has similar (or worse) dual monitor problems as you suggest above, it has similar (or worse) lack of software as you suggest above.
I only mentioned the items I have personally had issues with. If these are in fact "troll subjects", there's probably a good reason for it.
>> OS X, in the context of local governments trying to save money, is - and I say this as politely as I can - a stupid suggestion. The hardware is much more expensive, is a lot more desirable to thieves, it has similar (or worse) dual monitor problems as you suggest above, it has similar (or worse) lack of software as you suggest above.
Maybe you're upset because you are purposely mis-reading what I wrote. My choice of OSX over Linux specifically referred to enthusiasts, and had absolutely nothing to do with the Munich local government.
Note this is largely procedural. Employee can play with a spreadsheet 'for free', but a database requires finding budget for Business Analyst, DBA, Java Programmer, etc.
In many cases, Excel or Access functions as the prototyping tool for the ultimate real application.
I just today discussed with my boss a proposal from some finance guys of a "dashboard" in power point that will take half day to be manually updated every month. Come on... nobody sees this is plain wrong?
SharePoint is a CMS mixed with Document Management Systems with a sprinkle of Workflows plus a platform for small (but limited features) web-apps.
Some companies chose SharePoint because of the features and the idea of avoiding "NIH" or writing web-apps from scratch.
I think this is where the "hackers" (or geeks, whatever, hackers aren't hackers anymore these days) mindset don't jelled well with internal IT systems: we all want to develop something from scratch the way we want it, the way we know, the way we hope it will be.
Once your users are asking features of Excel to be implemented in your web-apps that use Ext-JS GRID as front-end you'll know what you're signing for...(and at that point pulling a 37signals makes you look...let's just say less useful...)
Didn't know of the web apps in SP, would like to see some examples although, don't trust it just yet...
Sharepoint also does Workflow, good for companies that have red-tapes.
Not everybody have red-tapes but for some that do have them, Sharepoint is there to red-taped-taped you :D
Oh and excel in the browser: filters, pivots, formula, sorting, etc.
Also, Excel for data entry! (Excel to Sharepoint List).
Don't forget InfoPath forms to fill out expenses!
SharePoint also solves the problem (depending on how you look at it) with business people not wanting to use Subversion or version control system. It has "Check in" and "Check out" features.
At the end of the day, you can't replicate Sharepoint features with scrappy web-apps. I know it's enticing, but it's a very very super hard sell because the numbers don't add up.
I don't know, in finance at my company with put huge ammounts of man hours in crafting, correcting, dealing, with tons of excel files. There are analysts that spend most of the time putting data in Excel sheets from the data warehouse system and shipping in outlook or uploading on SP. Is this cost effective? I doubt it. It is also an horrible way of working, you have no programmatic/easy way of comparing data: start trying to understand where a certain variation from the forecast, from a certain account, came from out of 30+ entities. It just drives people mad.
I think that some way in the middle is the solution...
I've introduced a number of people now to OpenOffice and they've much preferred it as it has given them back a somewhat familiar interface, after the widely-despised change to ribbon (a UI I actually prefer but it was a shock to many).
Word is still better than swriter, but not by a margin that I care about, and the fact that I have one-click PDF generation by default is a pretty killer feature, in my eyes.
Moreover, it's more likely that the GUI can be forced to be stable on Linux, while the underlying system receives security updates, because of the separation of components.
 Right now, the keystrokes I know for Microsoft Excel still work, but they bear no relationship to the UI at all.
The only problem so far is - once a year they need to use free windows-only application, that doesn't run in WINE, to calculate taxes. They just go to neighborns for a few hours.
My sister also has Kubuntu on her laptop, but she knows how to update it, so I don't know which version she runs now :)
If your relatives only use computer to browse web and write simple documents - linux requires less maintanance. At least that's my experience.
If you're in the US, I've been using the web app tax software and it's been great. Nothing to install locally and my progress is always saved. Maybe not for you if you are in a complex tax situation, but if your life is simple doing it online is incredibly easy.
VAT statements in Germany are a bit riskier though: If you give the Revenue Service permission to withdraw money and I file a VAT statement for you which shows a large income they'll just go and collect 20%. It's a bit an edge case but that's the reason I manually send the VAT collected each month. I'd feel much safer if the interface was authenticated in some way.
I did a calculation in 2006, comparing a Lenovo T60p to a MacBook Pro, it was roughly the same. Things did not change since then to my knowledge.
Never heard that in a discussion. I thought it was all about comparing like to like. The OP didn't say "for reading emails". Especially as he gave the example of 100% Office compatibility (Office 2011?).
Or you could compare live time value. I've broken several Thinkpads in a small amount of time, my wife uses a MacBook for more than 5 years with the newest operating system and applications.
Or, if you need email, only use a phone. Or if you only surf the web, buy a tablet. Or if you only do small amounts of web and email, don't buy anything at all and go to a free library. It depends.
But the argument was that a Mac laptop is much more expensive than a PC laptop for working with Office. Which it is not to my knowledge.
If you don't need it, don't buy it - no argument there.
If you're willing to cut back on the quality and capabilities there are very cheap PC options, but no very cheap Mac options. For someone who just wants to browse the web and edit occasional documents there is a lot of cutting back that can be down.
No. The Mac is cheaper in the end.
I also wanted her to be safe from viruses and malware. I think her time is well worth the price difference.
Linux is definitely cheaper than Apple.
Also need to make sure to explain about file formats, or chances are pretty good that they'll try to send a LibreOffice file to a MS Office user.
I remember Office 2007's ODT support being rather poor, too. Most documents I tried to put through it only came through with minimal formatting preserved.
2007 did do fairly poorly from what I remember as well. 2010 though is much better; I go back and forth between those fairly regularly. Fonts are obviously an issue, and complex formatting isn't quite right all of the time, but each suite seems to open most of the other's file types fairly well.
> Until 2003, the city was using Microsoft Windows NT 4 across the board, and was by and large satisfied. When Microsoft decided to end the support for this operating system, this meant that hardware and important procedures would eventually stop working. “It was from this experience of being totally at the mercy of an external party that we wanted to take the road to more independence”, Schießl (ed. the deputy project manager) says.
So having seen the effect of a vendor lock-in, the conscious decision was made to move to an environment where vendor lock-in can be avoided. It may take considerable effort to support a linux distribution after its official EOL but it certainly can be done - unlike any other closed-source OS.
The Munich migration is pretty much the canonical example of how to botch a Linux migration.
They "deduplicated" a lot of redundant forms (one little example: The act of asking for PTO was implemented in 45(!) different forms depending where you work in Munich - now it is one single unified service for all) and by centralising stuff they can now do useful statistics that simply didn't exist before.
The focus on Open Standards starts saving a lot of time and money in the archives already.
(Disclaimer: I work for Red Hat, have been involved privately in the migration a bit since 2005)
Does anyone think that large organisations will tend to move to Web-browser based business applications over the next few years?
If so, will that make it easier to switch the client device without major change in the central plumbing?
Yes. We're upgrading (slowly but certainly) JDE and it gets rid of all the fat / skinny clients in favor of Web. Which has been the trend in ERP for a few years now.
I don't know if it's a good idea or not - people (coughmanagement*) seem to underestimate the amount of resources ops needs to maintain large farms of web servers in the enterprise.
On the other hand, for a few years I was responsible for a bar-code server, which had it's own client to install and that was a white-line nightmare, let me tell you. I don't miss it at all.
Its a fun read to learn about how many instances of Firefox you can run on a single machine at any given time as well..
They are now on 10.04.
1. Completely anonymous - who is this person who has dedicated a site just to this topic?
2. Clearly neither German or even speaks German (using Google translation!) - so, we can rule out an insider or even a user.
3. Railing on all topics relating to Linux for government use anywhere on the globe without giving reasons other than quoting other sources.
As I said, not reliable and someone clearly disgruntled with no independent information or expertise (technical or managerial).
Sounds like a disgruntled vendor.
Either way, a 9 year IT project is really dismal.
Measuring the success of an IT project in years is wrong in any case. Success is defined in meeting your goals at the required time. Munich is a little late and they have had their own share of issues and roadblocks, but all in all they met their goals. Considering that this was the first large scale migration of a city administration in germany, that's not dismal.
And perhaps you would work through this sort of multiyear migration plan, but I certainly wouldn't. From what I can tell the project was driven by feel-goodisms and not at all properly scoped.
Added to that is the fact that the migration happened department by department in multiple steps, one done, next started. So yes, it's a multiyear plan and I do agree, I would not have planned and gone through that plan but calling it a failure because someone actually had the guts and vision and pulled it off is a bit far fetched.
Does anyone know details of this? Any ideas on what type of "optimisation" was done?
I recommend reading this: http://joinup.ec.europa.eu/software/studies/declaration-inde...
I am all for open source, and cities like Munich getting into it is welcome news, and will influence improvement in apps, however I suspect there is a lot of negativity pushed under the rug here.
(Also, the maximum number of complaints is irrelevant, but I guess complaining here about people not understanding statistics is pointless.)
"Ude said it was impossible to be exact about the amount of complaints the help desk gets about LiMux, noting that most problems are a combination of several causes. The software is not always the problem, since often there are problems reaching a server, or Internet connections might be malfunctioning."
My concern here is that they aren't logging incidents via some sort of appropriate iTSM framework (MOF, ITILv3, etc.) Even the most basic Incident Management setups would allow them to perform basic analysis of the incident data to work out where there issues are coming from.
I'm afraid I just don't buy the argument that it's impossible to know for certain where the city's problems are coming from :(