5 ms·
It’s because it’s not really doable without tremendous effort. I work for a muniplavity in Denmark, and we’re a big participants in an open source group for pub
by sidstling 8y ago
It’s because it’s not really doable without tremendous effort. I work for a muniplavity in Denmark, and we’re a big participants in an open source group for public software called OS2, we’re also members of a group of municipalities that own our own ESDH software and codebase and pay private companies to support and maintain it. Our libraries operate on 100% open source with Ubuntu supported by a local company.
So we’re actually pretty progressive in terms of pushing open source while also supporting local business.
We operate more than 300 IT systems though, I’d say that around 75% of these run on windows only, many without suitable alternatives. So even on the technical side, we can’t swap our stack because our employees wouldn’t be able to do their jobs. We’re working to lower this, and stuff like web-apps and the rise of android/iOS devises has helped but it’ll probably take 25 years to happen, and here’s the thing, Azure is actually the most EU friendly, secure, stable and cost-efficient cloud platform, so a lot of those non-Microsoft software actually still run on Microsoft.
Then there is the employee training. I’m not sure if you’re aware, but the primary cost of running a public organization is your employees. They are the most important resource and replacing them is really expensive. So is retraining them, and it’s also something we already struggle with in terms of IT.
Most times when we switch a system, the technical implementation will go excellent. The organizational implementation will go horrible, however, because learning how to use an IT system is hard. It’s even harder to learn how to use it efficiently and often local management will be reluctant to invest enough time or focus on what they view as IT changes, because they are already understaffed in their primary function.
Windows, android and IOS are something most people have worked with before. So is the office package, and again, where is the non-Microsoft alternative to the office365 enterprise stack? Anyway, switching people who can barely “turn on the internet” (talking of course about the browser, but that’s what you’ll literally hear every day in our support center) to open source alternatives is a tremendous effort that nobody, nobody, outside of tech wants to do.
Then there is IT, we’ve run Microsoft for decades. Our staff is trained for it, they’re certified in it and despite what you may think, they’re actually cutting edge in terms of skill. At least in my country they are, we’ve yet to find an ADFS consultant that knows more about ADFS than our guys, in fact, we’ve had to send most of them home because they weren’t able to help us. Where do we find Linux replacements for those people? And how do we pay the 25 million it would cost us to replace the ones that don’t want to work with the Linux stack?
Lastly there is support. Microsoft may have a spotty reputation in popular culture, but their support has always been top of the class to the public sector. If we file a primary incident report to Microsoft, Seattle will be on the phone with us until it’s fixed. What Linux house will offer this service? Even if we found one, I don’t think we’d really want to throw away decades of partnership with an untested entity just because.
I wish we could run European based open source software, I really do, but the truth is that it’s complicated, highly improbable and the business case isn’t even there. Because it would be really expensive, and do you really want to sacrifice welfare to run Linux? Would your political leadership?
- adrianN 8y agoI realize that it's extremely difficult to switch software systems that have grown over decades to something completely different and you can't do that over night. But then I see cases like the Munich project where they have been using Linux for several years[1], saving 11 million Euros in the process, until Microsoft moves their HQ to Munich and suddenly the council decides to switch back to Windows. In Germany at least there is tremendous potential for cost savings if it were possible to develop one FOSS set of software for municipalities and share it across the country instead of paying millions in licensing costs. [1] https://en.wikipedia.org/wiki/LiMux https://en.wikipedia.org/wiki/LiMux
- sidstling 8y agoI’ll admit that I’m not knowledgeable about the German public sector, but I’d really love to see how they managed to save money. I work primarily with digitization and efficiency, so I’m cursed in that I’m always thinking about cost efficiency. This is actually the primary drive behind my personal open source idealism. Since open source is customizable, we can cut away the clutter and focus on user experience and results. But because of my background I can’t help but think about the man-hours they must have spent on the organizational implementation, and how many hundreds of millions of euro that must have cost them. Maybe the German public sector is less digitized than we are, making the change lighter. Don’t get me wrong, we could certainly save money on licenses too. It’s just that the cost of doing it far exceeds the benefit, but I’d love to learn from Munich. The cost of change is why I personally think hard switches (I did notice them calling it gradual on the wiki page, but I’d argue that it’s not if you swap OS before every other system is ready) are a mistake in enterprise. We’re gradually changing our plumbing to be more open, but we’re doing it so that nobody notices. Our MSSQL cluster is a good example, it’s currently necessary because of legacy software. In one or two license cycles I fully expect it to be replaced by open source however. If things go as planned then no-one on the business side is going to notice anything but the freed up funds.
- adrianN 8y agoI would already be happy if all tenders paid for with public money would include a clause that forbids dependencies on proprietary software. There is no good reason to write new software that doesn't run on Linux.