Munich signs off on Open Source project The German city of Munich has declared that its famous move to open source software is over and a success. Munich famously decided to go open source back in 2003, citing a desire to be independent of big, bad, vendor-land and save a few Euros along the way. To that end the city decided to adopt Linux on the desktop and server …
It takes me an average of 6 years to move a company 85% of Microsoft and 10 years to hit 100%. A decade seems reasonable. It takes time to recode applications to standard or find/wait for applications that third parties are making to to replace ones that have no open source alternative.
But the transitions get made, and the savings are considerable. Enough to hire full time developers (even in SMBs!) to contribute back to the community by taking the applications coded internally and open sourcing them.
Every transition away from Microsoft that I've done helps to serve as an enabler for others. Maybe one day so many of these will have been done that a critical mass is reached and the bulk of businesses start walkung away from the chains.
A decade's worth?
10 years ago, OpenOffice struggled with many Microsoft Office documents. Microsoft's Office Open XML standard didn't exist, OpenDocument was not ratified as a standard, Ubuntu didn't exist, and hardware was at times, a hit and miss affair.
A lot has changed. That decade would have consisted of many pilot programs with various Linux distributions, numerous software package trials and probably many failures too. The key point is that they now believe they have succeeded, and I wish them well. I think we'll all be better off for it.
If you look at the state of Linux in 2003 vs it today it's a different beast. Hardware support is now excellent on the whole and Intel has now committed to not blindly following Microsoft for support. I.e. even better Linux support.
Also the figure of 6 desktops per day is ridiculous. Id imagine the early part of the project was about looking at the estate as it stood. Finding the alternatives that work over a number of pilot projects.
Anyway, it proves it can be done and I'd imagine the long term savings of this will be seen over the next 10 years.
Our government would do well to look at this study and work out if it would pay to get out of bed with Microsoft and stop signing these stupid deals.
And if all its going to set us back is 30 million then that's an investment for the future that will pay far bigger dividends than a new NHS or W&P system. (And far cheaper)
Well OK but if you had a clue of any sort you'd know that even in a tiny organisation of say 500 desktops, every single user uses their kit in slightly different ways, some will pull some obscure function in an Excel spreadsheet, some will have about a thousand Word macros, how many intranet servers with thousands of pages that all need checking and testing? Sorry but you cannot dictate that user's will do XYZ and lump it, this is the real world and when top notch manager pulls in 25% of the company business says "He hates XYZ new fangled desktop!", the IT troops scramble to find out why, get it sorted and all while still trying to keep it all safe, secure and manageable.
Seems obvious you've never done a desktop rollout in your life! Took my shop months to move up from XP to Win7 on only several hundred desktops 'cos we had so many different apps that all had to work perfectly else you get very high profile manager ( and their staff ) beating down the IT head's door and demanding a rollback.
I don't want to even contemplate switching operating systems on 15,000 desktops, that's a project manager's dream or nightmare, depending upon how it goes..
as only one of thousands/millions of small and large sites that are based on opensource/Free software. What is exceptional is that it is a high profile migration that has plodded on despite the large amounts of money spent by Microsoft marketing in a determined derailment effort.
When they started it was brave to the point of recklessness (for a large business environment with a number of windows-only legacy apps).
In this day and age it is not particularly difficult. In fact, there are lots of places where linux has displaced windows. For example are quite as likely to find xfce4 on a trading desk as windows nowdays - it is stable, requires little or no support, it is easier to comply with the farious mandates on it being patched up-to-date. Key apps are bespoke anyway and written in house so no difference in terms of app investment either.
Trading aside (as it has always been a special case), bespoke apps for anything above an SME are either written in java or in-browser as web apps and operate versus a specific backend environment like SAP. So they can run on Linux with no problem. In fact, they probably run better on Linux. This leaves only documents. While import/export from libreoffice leaves a lot to be desired (it has gone worse lately), if you stick to its native formats it is fit for most purposes. It is probably easier to use compared to MS Office as well (give the idiotic ribbon UI to someone who has been using legacy for 10+ years and watch the fireworks).
Someone starting such a project today will have a much easier time - it should be doable in a few months, not years.
I largely agree, but don't overdo it, in a large organization you do absolutely nothing in a few months. It's harder than that but doable, yes. Having worked for more than ten years in a +10.000 company I was annoyed by how that company probably kept on paying all kinds of royalties, year after year, on software that was probably scrapped years ago. It would not surprise me if Munich has saved some money getting rid of unnecessary royalties during this project as they have been forced to look at it in detail.
"And then
An excel spreadsheet full of macros comes in by email to the senior auditor."
They've already covered that in their documentation :
http://www.muenchen.de/rathaus/dms/Home/Stadtverwaltung/Direktorium/Strategische-IT-Projekte/LiMux/Dokumente/Praesentation_LiMux_engl_web.pdf page 13-15
> How does that solve the problem?
Munich is a customer of the auditors. Outside of the MS/Oracle/Apple bubble, customers actually get to tell suppliers what they should do if they want to keep on being their suppliers.
i.e. If you want our business, you must communicate using the file formats that we specify.
An organisation that has chosen a MS platform has just as much right to specify that suppliers use MS formats. Why should Munich be any different?
> So to correct the very misleading headline figures above, in reality, after ten years, a full 30% of their user base still needs to use Windows.....either physically or via Citrix...
Windows is irrelevant. Some of the user base needs to _sometimes_ access specific _applications_, which are probably third party. It happens that those applications still only run on Windows.
You seem to think that they have done this because they hate Microsoft. Actually they have done the conversion first to save money (which they have) and also to have control over their own systems rather than having other companies control them.
You may hate 'the opposition', this is obvious from your posts, but others aren't driven by hatred. You may generate FUD (nothing you say is actually true) but, while many have a MS policy due to fear, others aren't intimidated by your rants.
But the real question is: why do you care with so much passion and hatred ? Does every Linux installation diminish you personally, do you lose yet another brain cell every time someone does not buy a copy of Windows ?
"Windows is irrelevant. Some of the user base needs to _sometimes_ access specific _applications_, "
So they need to run, maintain, support and license Windows for 30% of their user base. That's not irrelevant. That shows why Linux is still irrelevant on the desktop.
" Actually they have done the conversion first to save money (which they have)"
No - it has cost them over €30 million MORE than to upgrade their existing systems. The claimed 'savings' have been widely debunked.
To suggest that they can build a new OS, run a migration project for TEN YEARS! - and support 2 environments instead of one across 30% of their users and then save money is clearly not credible.
Oh FFS...
So they need to run, maintain, support and license Windows for 30% of their user base. That's not irrelevant. That shows why Linux is still irrelevant on the desktop.
No, that shows exactly why suffering with vendor lock in is a bad thing - they have left over (3rd party) proprietary software that only works on Windows. In an organisation of this size there will always be some cruddy software usually supporting specific closed hardware that only runs on Windows. Often this is because the supplier has no clue about anything other than Windows development or has no interest in re-developing for other environments for low volume customers as it just doesn't make financial sense for them (the supplier of these proprietary systems) and from experience it usually doesn't seem to make financial sense for the supplier to ever update their software or to make it usable either... The more that open communication standards and similar are used the less this is a problem as there are more alternatives.
No - it has cost them over €30 million MORE than to upgrade their existing systems. The claimed 'savings' have been widely debunked.
Wrong. It's your pay-master's lies that have been widely debunked. By Munich themselves , not by third party interested parties or spreaders of FUD. While it can be argued that Munich will put whatever spin they want to on these things, the reality is that as a public organisation they operate with a level of accountability and if they did lie or misrepresent the truth it would come out very quickly and they'd be held accountable for it.
To suggest that they can build a new OS, run a migration project for TEN YEARS! - and support 2 environments instead of one across 30% of their users and then save money is clearly not credible.
Wrong. Again. When you are forced to support Windows you usually have about seven different versions of Windows and Office to support, all incompatible with each other in all so many delightful ways requiring specific customisations and management of each independently. In the end you typically wind up with seven different configurations and when you make a change to one you then have to reflect it manually in each of the others as appropriate... and that's just the joys of centralised management aka group policy. This is before you start to take into account the various incompatible and always fundamentally broken or deficient security structures that are in place that need to be realised between sharepoint, exchange/outlook, file shares (different versions of course), device (e.g. printer) access let alone network access.
When you have a budget the size of the city of Munich and you ask Accenture/KPMG/Whoever to jump the answer is immediately "How High".
They all have their own IT shops so they will solve that problem if needed to keep the contract. They will probably re-sell the know-how against IBM in other tenders after that too.
"Good luck finding a replacement auditor that doesn't rely on Excel"
They are the eighth largest metropolitan economy in Europe. They don't have to go looking, auditors will come running.
Besides them being, you know, the customer who the auditor should be trying to accommodate. You think that Open Office Calc has no way of getting data from Excel?
"Good luck finding a replacement auditor that doesn't rely on Excel.....You will be hunting for a long time!"
In Munich?
Really?
And you couldn't get the document to opne as or save as in Office Libre or OOo?
Or get the sender to convert it to ...?
You would definitely be using the wrong sender. Pity it is likely to be a home computer user but luckily the only people sending you macros from those sources are going to be script kiddies.
"Just when the rest of the planet is moving to the Cloud for a better ROI". Himalayaman
Given that the clients are moving off the traditional PC, the software company can't keep selling them the same desktop client over and over again. Moving them to the Cloud and they can charge them software rent into perpetuity.
"Moving them to the Cloud and they can charge them software rent into perpetuity."
Shhhhhh. That's not the message the cloud-boosters want people to hear (though Trevor and others do seem to have spotted it).
On the other hand. cloud-style internal shared systems, properly designed and operated, with lots of zero-Microsoft desktops ... yeah man, that could be good for lots of organisations. Except the certified Microsoft dependent ones.
While undoubtedly an achievement to move such a large environment - especially one which by all accounts was historically dreadfully maintained - over to Linux, I can't help thinking that this shouldn't be the poster boy for Linux/FOSS rollouts. If I were to convince a CIO of a company to move over to Linux/FOSS from proprietary software I would not hold up a decade long project as an example of how things can go well. Especially when it's seriously questionable if it saved them any money at all. I've been following this for ten years and initially there were lots of comments about how much money was going to be saved, these have all gone now and been replaced with comments about freedom and flexibility. They don't even have support for their OS, it's all provided internally, which would be something totally unacceptable for the vast majority of companies.
"They don't even have support for their OS, it's all provided internally, which would be something totally unacceptable for the vast majority of companies."
Internally provided high level support is unlikely to be acceptable when the OS and apps are closed source with a commercial ecosystem with decades of experience of providing for-profit "support" (ie do it the MS way or take the high road).
Internally provided support is a different matter when the source is available. Small and many medium organisations may still want to farm out their support, but some (especially larger ones) will also have the option of bringing IT support in-house.
sorry, but I think you misunderstand my point: The majority of companies won't accept no support, proprietary or FOSS, and they won't want to employ kernel programmers to provide in house support because it's way outside of most companies core business. I would not be able to take a proposal to board level which went along the lines of "we'll move over to X system, first off we'll need to employ a load of programmers in order that we can support and maintain internally." because the reply would be "but we're a bank/florist/heavy engineering company. We don't do OS programming."
"I think you misunderstand my point: The majority of companies won't accept no support, proprietary or FOSS, and they won't want to employ kernel programmers to provide in house support because it's way outside of most companies core business."
"we're a bank/florist/heavy engineering company. We don't do [critical but non-core stuff]"
Should these companies employ their own electricians, plumbers, etc, or should they outsource it as required to a trustworthy supplier?
Answer: they shouldn't care, as long as the job is done properly. Why is IT support any different? And with open source you can choose who supplies which pieces, and when you upgrade it's on your schedule rather than being dictated by your suppliers. Just like you get with the electricians and plumbers - be they in-house or external.
ps
lots of banks employ lots of IT people already. If they need DIY OS people to help run the business, why not?
Employing an external company to carry out your non-core OS maintenance IS support. That's my point.
And, yes, banks have lots of programmers, but those programmers are banking IT programmers, not FOSS OS programmers. They also already have a job, you'd still need to employ more non-core skills on the off-chance that they were needed.
Running a system FOSS or COTS without support for most companies is just a no go area, Munich don't operate with support, they use their build team to carry out the work required, this would not be considered acceptable to any company I've ever worked for.
"you'd still need to employ more non-core skills on the off-chance that they were needed."
Now who's misunderstanding.
You could do (some of) the work in-house, or you could farm (some of) it out to a "shared services" company. Much the same as many businesses already do with plumbers, electricians, and similar "facilities" work. This is stuff that's critical to the business, often highly skilled work, but not always "core business activity".
Please explain to readers why IT is *that* different,
Obviously IT is different if only MS have the OS source and therefore only MS or MS-authorized organisations are able to do sensible fault-fix or upgrade work, and if all kinds of impenetrable commercial agreements dictate who's allowed to sell what to who, and when (just ask Trevor about MS licensing). Then IT is trickier than plumbing.
FOSS can change that, either with in-company support or with shared-resource support.
Choice is good, right? It's what fair markets are about, right?
But choice is only meaningful if it's informed based on relatively complete relatively unbiased information.
"lots of banks employ lots of IT people already. If they need DIY OS people to help run the business, why not?"
Because it makes far more sense 99% of the time to buy something off the shelf so that those maintenance and R&D costs are shared across many other customers...
"up a decade long project"
Although a decade is often mentioned it looks as though although the decision was made ~2003 actual implementation as opposed to pilots didn't begin until ~~2007, ran as a rolling process whilst maintaining the full council services and indeed had been intended to take quite a time.
"f what you claim is correct, that is frankly utterly staggering."
Why ?
I've seen the time-line on a Munich doc but can't find it at the moment but from memory the bulk of the desktop moves took place rather recently (~3 years ?)
Yes it's on the document :
http://www.muenchen.de/rathaus/dms/Home/Stadtverwaltung/Direktorium/Strategische-IT-Projekte/LiMux/Dokumente/Praesentation_LiMux_engl_web.pdf page 6
Looks like they spent 2 years getting people onto OO, then 2 years pilots of the linux system (1), then roll-out from 2011.
(1)Until the end of 2008, each of the city's departments will have a "LiMux germ cell". These are groups of 30-50 workstations that will be migrated to the LiMux client. Even in departments that are sceptical towards the migration, this helps the IT staff to become familiar with the software. This approach also allows the LiMux project team to learn about the specific technical requirements of each department, and address them before the full-scale roll-out of the software.
"Why ?"
Because it is a horrifically long time - to end up with what is effectively a crippled desktop compared to a current Windows based solution - and they still can't support 30% of their apps on it!..
I ran a programme that migrated 70,000 end user devices to a new OS in less than 2 years on over 100 sites globally including rationalising and repackaging ~ 8,000 client applications. The people running the Munich project must be laughing all the way to the bank to stretch it out for a decade and still not be finished!
> I ran a programme that migrated 70,000 end user devices to a new OS in less than 2 years on over 100 sites globally including rationalising and repackaging ~ 8,000 client applications.
We have no way of validating that claim. Though it does seem rather dubious that such a big cheese would be wasting their Monday afternoon posting as an AC on the El Reg forums.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
