Should be an EU wide mandate at this point, as it's mostly pros with little cons.
Since the options are 1), keep spending EU taxpayer money to boost Microsoft's and other US big-tech market cap, or 2), spend EU taxpayer money on FOSS and local companies implementing and maintaining that FOSS for local infrastructure, then the correct choice seems obvious to me.
The only cons are the short term costs, teething issues and pains of the transition, but that's outweigh by the long term pros once that hill is crossed, the biggest which is tech sovereignty and independence form major tech firms under Uncle Sam's control, and the taxpayer money going to local companies and jobs instead of US.
Ironically, Russia was unaffected by the Crowdstrike incident since they're under sanctions and can't use it, and IIRC they also started switching to Linux for their infrastructure after the war, so maybe it's an interesting case study.
The biggest con is IMO the most obvious one. What if a third party non-open source product does something that FOSS can’t do or is dramatically behind on?
And are companies going really source local maintainers for FOSS versus the standard maintainer? Eg, is Switzerland really not going to use Canonical for supporting Ubuntu because it is British?
> What if a third party non-open source product does something that FOSS can’t do or is dramatically behind on?
That's almost the best side effect: in a place where open source is mandatory, they will have to fund the missing features and it will be available to everyone, once and for all. If everyone does that, we rapidly get great software, sharing the costs, very efficiently, and it's all open source.
The alternative being everyone paying their own little license on their side ad vitam eternam.
In theory, maybe. I think in practise, in the case of somewhere like Germany, they realistically just won't implement the hypothetical feature and let productivity of employees drop even further. There is no motivation or reward in the public sector here for anyone to do anything better or more efficiently. I'm maybe overly cynical about the public sector's ability to get anything done here, I would like to be proven wrong, though.
It seems like maybe you’re assuming that state employees would be the ones building the required features, but I would guess these are simply going to be makework contracts going to (hopefully) local engineering firms.
No. I’m not assuming that. I assume the governments will have to fund the work, but they will contract it out. But even then governments have struggled to deliver on this.
And even when governments have the funding they tend to be horrible at specifying and getting delivery of technology - at least that’s been the case for the US.
> And even when governments have the funding they tend to be horrible at specifying and getting delivery of technology - at least that’s been the case for the US.
This is the dysfunctional contracting model combined with “cost savings” removing technical civil service positions, meaning that the government often lacks people who are qualified to review or manage contractors. Groups like the US digital service, 18F, etc. which have staff have much better track records.
Funny. I thought “ok, yet another generation of Accenture/TCS/Fujitsu large scale government projects that cost billions and fail to deliver a single thing”
> in a place where open source is mandatory, they will have to fund the missing features and it will be available to everyone, once and for all
Or the majority of their users will suffer with substandard products, while highly paid or influential users will get exceptions to the policy so they can use the software they want.
How is it relevant to open source? Today, users already suffer with whatever substandard product was chosen and enforced. Except it's also probably proprietary.
This idea works for well funded open source projects - typically funded by Big Tech. But for a given country to do this it seems more likely the functionality is just missed until Big Tech decides they need it.
For example let’s take apiece if software everyone on HN hates, Sharepoint. I’ve yet to find any open source alternative that comes close. And while you may hate it, it provides tons of value to local governments I’ve worked with.
Reading the blog linked in TFA, in this case the requirement applies just to software developed by the Swiss government (or on their behalf by a contractor), not to all software in general.
I don't see this as a problem, at least not more than it already is even with proprietary software. Remember that announcement from just a few years ago that Japan will finally move away from using floppy disks as the official way to exchange data with some authorities? Or stories about government offices somewhere struggling to move away from windows 7?
I really don't see how it even still matters today. Even a Libre office from 5 years ago should have plenty features for the average tasks an office clerk needs to do. Heck probably even MS Word 97 would suffice. The rest is often specialized software written by contractors, so "make it OSS" will just be a new requirement.
In theory, of course not. But in practice sure. Simple, very accessible example, there’s no Madden open source alternative. And I’ve actually seen people try to do this.
I’d love to see an open source football/basketball/sports engine. But it’s hard to do, even without dealing with the licensing aspects.
I think a big part of the reason is that Big Tech really helps push along FOSS, but the licensing aspects make the ROI poor for Big Tech to invest in FOSS here.
That's nonsense. Most of the software in question was written to support the administration. There is no existing closed source nor open source software as such, only software ordered by public sector and delivered by private sector under a license that guarantees vendor lock-in. Which, obviously, is insane policy and must go.
The software is updated to contemporary standards and rewritten as people come and go about every 10-20 years. Mandating it to be open source makes it simpler for new vendor to take over in those 10 years. And it might provide other benefits such as some degree of code reuse.
As for the foundational software, there usually is a regional partner of Canonical, SUSE, Red Hat, Microsoft and so on who re-sells the licenses and provides part of the support.
Isn't the Linux/GNU philosophy "Use the best tool for the job". Making OSS always a choice is good, but it is definitely not always the best tool for the job.
Even if it's not the best today the question is for tomorrow.
With proper funding you can hope open-source will become better and hopefully if more actors join the cost might even become lower.
Of-course you can continue to enrich the likes of Microsoft today because they are "better" but even not considering the cost with the current trends in Microsoft (and others) products you might still end-up with something worse and another expensive bill to pay to try to leave ...
“Best” isn’t a simple objective trait. Is that the one which lets you get your job done for the lowest cost? Or the one with the easiest training and documentation? The most consistent UI so you don’t have to retrain people? The one which avoids geopolitical concerns about ceding control to potentially hostile countries? Is it the one where you have the most control over the direction because e.g. there could be an Swiss digital service cranking out patches for things they would have to haggle with someone at Microsoft or Oracle to get implemented? What about the one which you can most easily fix accessibility problems with? (That last being a legal requirement doesn’t mean that certain companies don’t try to blow it off release after release)
Every one of those will have someone arguing that it should determine “best” status. A compromise like “usable and we have more control” seems quite defensible.
Sure, but if you let most people choose, they don't choose the best IT tool. They choose what they know. Or they choose the tool with the best marketing team.
Either way, you can't make a rule for people to choose the best tool for the job. A rule like this may come closer to the GNU Linux philosophy.
The much bigger and more obvious con is interfacing with "everybody else" (e.g. outside the public sector). People assume perfect bug-for-bug compatibility with the dominant software solution. So you also incur long term costs of always ensuring that compatibility.
Hence, the more common compromise is to pay a local company to implement and maintain that infrastructure for you but using the US software.
> Since the options are 1), keep spending EU taxpayer money to boost Microsoft's and other US big-tech market cap, or 2), spend EU taxpayer money on FOSS and local companies implementing and maintaining that FOSS for local infrastructure, then the correct choice seems obvious to me.
Didn't that happen in Germany, where they basically did a migration away from Microsoft Office to LibreOffice and then basically switched back? Here's a more recent article about something similar that mentions that case: https://www.zdnet.com/article/german-state-ditches-microsoft...
> If some of this sounds familiar, congratulations on having a great memory. Munich, the capital of Bavaria, Germany, switched from Windows to Linux in 2004. That move lasted for a decade before Munich returned to Windows -- in no small part because the mayor wanted Microsoft to move its headquarters to Munich.
Either way, to me it feels like both having that additional bit of options/leverage is a nice thing, as well as using open source in the first place.
I've built systems both with the likes of Oracle and PostgreSQL and while both have some nice features, in most cases I'd lean towards the latter because it's just nicer to use, regardless of whether you need to operate an instance for some environment, launch a local instance or even have short lived test instances that are spun up for your CI pipeline; it's like suddenly your hands are no longer tied up in licensing and layers upon layers of complexity and restrictions. In my eyes, it's the same for using some variety of GNU/Linux on the server side, instead of Windows Server.
That's not to say that tech coming from big orgs itself is the problem, I have nothing against JDK, Java, MySQL, .NET, ASP.NET etc. Rather, it's that permissively licensed solutions with communities around them are usually pretty nice to use. Plus, there's way less friction during development, too., you don't even need to check whether you have the Kendo UI license but instead something reusable is an npm install away.
That said, in regards to established software like Microsoft Office vs LibreOffice, it's likely that there will never be full feature parity, but that's probably not some absolute dealbreaker for most.
>
That said, in regards to established software like Microsoft Office vs LibreOffice, it's likely that there will never be full feature parity, but that's probably not some absolute dealbreaker for most.
At least for finance and insurance industry, the feature differences between Excel and LibreOffice Calc are are deal-breaker (users in these industries are often power users of Excel). I can also easily imagine that the same holds for Powerpoint vs LibreOffice Impress for the consulting industry.
I'm always in awe on how petty and pedantic some people can be in the comments here, acting confused and pretending to misunderstand you out of context, if you aren't 100% verbose with your statements as if they're context unaware of current world events. They're worse than LLMs.
The same would apply to assistive technology. Right now, various countries are paying relatively big sums to Freedom Scientific for JAWS licenses. It would be way better for the users if those countries would pay open source devs to work on accessibility for everyone. However, capitalism is preventing this. Resellers would be out of bussiness, so they lobby for the current system to stay as it is.
Spoiler alert: it didn’t work well. The world is not perfect and decision makers are greedy.
Edit: I see I am being downvoted. Is there an objective reason for moving back to Microsoft? Is there any objective reason to ditch implemented and working open source software project? As Wikipedia tells most of city computers were successfully migrated. And users were giving positive feedback.
Seems like it wasn't so terrible. Munich has still decided that "Where it is technologically and financially possible", the city will emphasize use on open standards and free open-source licensed software. Which seems reasonable.
from [1]:
In 2018, journalistic group Investigate Europe released a video documentary via German public television network ARD that claimed that the majority of city workers were satisfied with the operating system, with council members insinuating that the reversal was a personally motivated decision by lord mayor Dieter Reiter. Reiter denied that he had initiated the reversal in gratitude for Microsoft moving its German headquarters from Unterschleißheim back to Munich.
local decision makers in particular. Doesn't take that much to get some administrators in Munich to flip and to run into compatibility issues when every city does it independently.
Make it a five or ten year EU plan with some serious cash behind it, it's basically at this point an investment into security and autonomy of the continent.
Of course Microsoft et-al won't sit around with their disks in their hands watching their market cap get obliterated by Europeans moving to FOSS, so they'll wine and dine politicians, open an office here and there, offer generous kickbacks and discounts for volume purchases, bribe Ursula v.d. Leyen and have her delete her emails after doing deals under the table, you know, the usual EU drill, making the decisions to move to FOSS seem very expensive to the bean counters running the show compared to staying locked in to the Microsoft, Google, Apple, etc ecosystems, so I don't think such a transition will even happen EU wide besides a few small municipalities here and there.
Sadly, this article is plain wrong.
I've been working in the gov. IT sector in switzerland and reading the headline and article gives a false impression.
The aforementioned law EMBAG does not mandate the use of open source software.
What it does is mandate that software develeoped especially for the federal government needs to be open sourced, unless that is not possible due to licensing issues or national security reasons.
The whole swiss administration is microsoft based.
Also, most day-to-day administrative tasks are handled by the cantons (think state) administrations, and they are, as far as I understand, not subject to this law. The federal and cantonal administration are quite separated in switzerland.
That's e-voting and that was an absolute disaster.
This is E-id which we voted on not to have the private sector operate.
The current solution proposed are quite good, the biggest issue that remains is that currently the law is not clear who can require the ID and this should be tightened so I don't need to show my ID to use a pay toilet.
Love the initiative but it has failed so often (Munich anyone?) because of stupid choices like:
- let’s make our own distro.
- Let’s use OpenOffice (0 knowledge of FOSS at the start)
- let’s ban anything proprietary from day one
I hope someone wise is in charge (that Sturmer seems to be a good one, let’s just hope he is not too dogmatic/RMS about FOSS) and they set themselves up for success. Which, to me means: take it slow, ask for feedback, guide; don’t force.
… basically apply standard change management techniques.
Well, yes and no. The law (which went into effect 2023 BTW) still has a lot of loopholes, e.g.:
> The federal authorities subject to this Act shall disclose the source code of software that they develop or have developed for the performance of their duties, unless the rights of third parties or security-related reasons would preclude or restrict this.
"unless the rights of third parties [...] would preclude or restrict this" pretty much says it all.
Somewhat ironically, Switzerland has four official languages, none of which are English, but with so many regional dialects within those, English is quickly becoming the… ahem… Lingua Franca:
You'd get by just fine. It's not uncommon to speak three languages fluently in CH (given it has 4 official languages), and usually English is in the mix. I lived there for a couple of years and while not speaking French or Italian, I got by quite easily in the respective regions.
Get by, yes its often enough. Job search - that limits your choices maybe to 10%, in a job market that is consistently overcrowded (100-500 applicants for any good position, not all are brilliant and with right experience but some always are, whole world tries to get here).
You will regularly hit the language wall in places like shops, post office, neighbors etc. where answer in given language if they speak english is consistently "no". Integration outside big cities is a lot about you becoming like them, not vice versa, better keep the differences compared to locals behind doors of your home. Smaller cities, more rural places - forget english outside tourism, many young speak it but otherwise there isn't much will and often neither skill. French part is worse than German part in this (just like French from France are much, much worse when it comes to will to speak english than Germans from Germany are).
Mandatory is a ridiculous rule if only for the self defeating nature of limiting yourself. I can understand the sentiment but maybe make the rule say to prefer OSS. That is when sourcing RFPs prefer OSS solutions that can fulfill requirements.
Since the options are 1), keep spending EU taxpayer money to boost Microsoft's and other US big-tech market cap, or 2), spend EU taxpayer money on FOSS and local companies implementing and maintaining that FOSS for local infrastructure, then the correct choice seems obvious to me.
The only cons are the short term costs, teething issues and pains of the transition, but that's outweigh by the long term pros once that hill is crossed, the biggest which is tech sovereignty and independence form major tech firms under Uncle Sam's control, and the taxpayer money going to local companies and jobs instead of US.
Ironically, Russia was unaffected by the Crowdstrike incident since they're under sanctions and can't use it, and IIRC they also started switching to Linux for their infrastructure after the war, so maybe it's an interesting case study.