Over the years, as they have continued to not actually migrate from from Windows to "LiMux" Linux, the city of Munich has moved the goalposts several times:
- In 2004, they said they would convert 100% of their workstations by 2008
- In 2006, the city admitted it would never be able to convert 100% of their machines, so changed the goal to 80% conversion by 2008.
- By 2008, with less than 10% of their machines migrated, the city promised to convert"a large part" of their machines by 2009.
- At the end of 2008, they promised "most of our workspaces by 2011"
(The sources of all these quotes can be found here.)
Today there is an article at Heise.de in which Munich project leader Florian Schiessl now pushes the date back to mid-2012:
Bis spätestens Mitte 2012 sollen 80 Prozent der 14.000 Computer in der Stadtverwaltung auf Linux umgestellt sein.
Google translation: "No later than mid-2012 to 80 percent of the 14,000 computers in the city administration to be Linux."
2012 is the new 2011.
Posts
16 comments:
Greetings from the Linux Hater's Blog. You're doing great work here. I'm surprised you aren't getting more comments--from either side.
Thanks, LHB. No one seems to care about the Munich vaporware project except me, but that's fine. I make rude comments to myself when I feel neglected.
Thank you
That's not true!! I check your blog at least twice a week. I'm interested in the Munich project, but I am a realist and thus very sceptic when Florian Schiessel's enthusiasm brushes over many things. Linux on the server - great - a cheap Unix. And that's it's greatest advantage, the non-existent (well, bar support subscriptions) licence cost which was the original motivation. It's called GNU is not Unix for a reason - It is a clone, a rip-off. There is nothing wrong with that, but all that "innovation" linux advocates pride themselves with was done in the commercial world. But that's still ok.
On the desktop, well performance is less crucial, what counts is eas of use and a large base of applications. That is the main shortcoming and it WILL NOT change if ABIs and APIs change ALL the time. And for a developer it is a most frightening situation, especially because there is no one at the helm. Take Apple and MS, you know your code is going to compile or your binary will run over different versions (ok Apple broke compatibility not too long ago, but that was managed very gracefully).
Linux is a mess and most of the work is done by commercial entities anyway, because they can have a Unix and not pay licence fees. Hence they do it for financial reasons and charge support contracts (without no CIO would touch it because of the risks of stuff breaking and no one is responsible for fixing it) - so free as in free beer has not workded. Also free speech has not as with commercial support you become dependent anyway.
Granted, Limux has its "own" distribution, but they will soon end up with a VERY old debian for which they have to write patches themselves. Spending time and thus money (tax payer money in the case of Munich) is CRAZY. This will become a black hole. Once they are (finally) finished deploying they will have to start an upgrade. They do not grasp the main problem in any desktop deployment - upgrades must be quick or it becomes a never ending story. And costs must be calculable. They will rise in Munich and not fall
they will soon end up with a VERY old debian for which they have to write patches themselves.
This is something that doesn't seem to be well understood: it's far cheaper to pay a software vendor for licenses and customizations than it is to become a software vendor (by maintaining your own Linux distribution). Unless you recoup development costs by selling the software at a profit, there's no way it will ever work out.
That's exactly what I meant earlier in my post. The Limux guys will find themselves using more and more manpower (== time == money) to "do it all themselves". There are situations in non-it companies where it makes sense to have some in-house development (Web CMS, ERP report customisation, etc.) but infrastructure software development is usually a no-go. And that's exactly what Limux is trying to do. I wish them the best of luck, but I doubt that even the (only) perceived gain of freedom can justify the cost in the long run. The greatest gain will stem from the business process and DB consolidation they did. I would have recommended doing that and staying with MS client side. However I would stay away from MS server side (apart from AD for client configuration). There are way cheaper options out there (Sharepoint and SQLServer are a rip-off)
infrastructure software development is usually a no-go.
Yeah, that's because after 40+ years of computer science and operating system development, there's little mystery as to how it should be done, and there are now off-the-shelf basic infrastructure solutions for every conceivable scenario that are designed and marketed by companies who are experts in the field. Glassy eyed idealistic government agencies can't compete with this.
No one's (successfully) starting new infrastructure projects these days. The home-rolled ones that still exist have been in place for a long time. Even then watch as these drop like flies as maintaining and extending the ancient COBOL systems becomes more expensive than a corporate wide infrastructure refresh.
I like your blog.
I like your blog, too. Everytime reading this shit, I can rofl for minutes ;-)
It's quite easy to shit on someone while being anonymous. So tell your correct name, adress and so one and we'll have one or two beers, answering your questions.
I wonder why the upgrades must take place ?
If you have a system that just works - why the upgrades ? Lot of upgrades going on now is Microsoft quitting support for older systems...
You might want to update a single app / patch bugs... but system wide upgrades could be a thing of the past for Munich
"I wonder why the upgrades must take place ?
If you have a system that just works - why the upgrades ? Lot of upgrades going on now is Microsoft quitting support for older systems..."
Because updated systems often bring productivity improvements? Oh wait, we don't need that! Hell, my university uses an infrastructure dating back to the 70s. It "just works", until you realize its slow as hell and a pain to maintain.
It's easy being anonymous because what he's saying has people looking for his address. I wouldn't want you showing up at my house, either.
Sorry, government projects have always been money-holes (at least in Germany). I do not think it has something to do with the technology involved.
There are many many examples of overspending and fucked up projects, and bare-to-none examples of government projects on time and budget.
I am webmaster of Itsolusenz Software Development Company Microsoft Certified, IT Solutions Company, offering Custom Software application development and web site development in India.
"Government projects have always been money-holes"
Don't mention Accenture. I did it once, but I think I got away with it.
PS: Why can't I copy/paste/move the cursor in this commentbox?
you have a nice site. thanks for sharing this enormous resources. keep it up. anyway, various kinds of ebooks are available here
http://feboook.blogspot.com
Post a Comment
Make excuses for Munich here: