▲ 48 points
back
60 comments
It always hits the nerve to hear that a public organization is moving back to MS from FOSS, but I wish this kind of demand(from governments) would lead to a open company that is strictly focused on developing an alternative to Windows desktop. Linux is heavily pulled in the server direction, and we need a company other than RedHAt who can pull Linux ecosystem a bit together to build an opinionated but coherent desktop.
Trying to support all of the email clients, calendar tools and obscure media decoders will never lead us to an usable Linux based Desktop.
If you consider that every municipality in EU (or world) has similiar problems and the costs for running Microsoft and other commercial software is quite high everywhere I really don't know why the EU doesn't start some public company or organisation and polish the existing open-source tools.
It's probably considered anti-competitive but let's image they scrap 1 billion of their budgets and invest in opensource:
- polish Samba to be a viable Windows Server replacement. Replication, missing Kerberos functionality, Implement ADWS, ADFS - maybe Exchange APIs. Save millions on server costs.
- unfuck LibreOffice, develop a strong foundation to implement typical processes
- polish KDE / Plasma to macOS/Windows usability/stability - rewrite kmail/korganizer with modern usability/ui
Not sure if that's possible but it feels like a lot of things are 60-80% done - maybe taking a leap and fixing the other 60-80% might make a difference.
- have an EU wide service agency that supports municipalities in implementing OSS
- demand opensource software for public money.
- create strong open standards and enforce them by law for public institutions.
maybe it will take 5 to 10 years but everyone should profit.
It's an unsubstantiated claim but my impression is that governments are not good at managing companies.
Who gets to decide if it's better to use LibreOffice or OpenOffice? What are the features that are the most important for the office replacement? How is it going to be supported? Is there a line where people can call to complain about the product? Taking all that OSS and packaging it sounds a lot like running a company, especially when multiple "consumers" are involved.
Good questions. Maybe something like a cooperative¹ that is independed of the states? Such a project could probably even be refinanced by users who pay for consulting.
How to organize this in such a way that it will be successfull and efficient? Let's start researching it now.
It's such a shame how European universities and every other part of the state is switching to Microsoft because the on-premise alternatives mostly suck.
I agree Linux requires some polish, but have you tried e.g. LibreOffice lately? It's very good.
I have a small deployment of Linux machines running GNOME and the typical office suite applications, including LibreOffice. I'm using NixOS to ensure everything runs extremely controlled versions of software. It's rock solid.
Furthermore, the users are really technologically illiterate. And they like the UI. They have no issues.
I think the big problem is usually that people expect a perfect replacement for MS Office. They try to open DOC (not DOCX files) in LibreOffice. And it's not perfect. It can't be perfect, as it's playing a very hard catch up game on ill-defined file formats.
> I think the big problem is usually that people expect a perfect replacement for MS Office.
Nobody owes it to LibreOffice to use it when a superior alternative exists. DOC support is minutiae - there are certainly people who do use it, but intentionally switching away from the default (DOCX) while Office moans at you about reduced functionality seems like a huge degree of dedication to a file format. Two huge reasons why LibreOffice is currently not good enough for me:
* The UI is functional and tolerable. It's not pleasant to stare at all day. Menus, toolbars and floating windows are fine if you hail from the 90s. Mozilla are a shining example to other OSS projects of what users expect in 2018[1]. Make it work, make it right, make it fast. OSS needs to learn that there is another step: make it beautiful. And they like the UI. Do they love it?
* Advanced integration with Dropbox or what-have-you. Microsoft has Sharepoint. So far as I know, Office is the only thick-client suite to offer collaborative editing on top of file sharing. When you have dozens of users collaborating on a project this stuff is incredibly important. Email is not scalable - your users shouldn't be opening a DOC from an attachment in the first place.
There are 13000 reasons why Lower Saxony is switching away from OSS. Dismissing what the competition offers is dismissing those 13000 reasons and means that a project will never address those reasons.
Good enough is not good enough.
FWIW, current releases of LibreOffice 6 can open and edit documents from SharePoint document libraries.
Also, I have not had trouble opening OOXML files (.docx, .xlsx) in LibreOffice.
The main problem with LibreOffice, from my point of view, is that there are many third-party applications that expect MS Office to be installed or use it exclusively for data import/export. For example, Autodesk Inventor can create a bi-directional link to Excel and export the list of parts of a model to Excel; ePlan, too. Our PDM system comes with an Outlook plugin that allows users to upload emails. Our ERP system can export stuff to Excel, but not to LibreOffice. The list goes on and on. For a large number of use cases, LibreOffice is perfectly cromulent as it is. But as long as all those third-party application rely on MS Office, it will remain dominant, regardless of the price or technical merits.
If you extrapolate a bit you can see that we can get rid of governments and have private companies run our lives completely.
Public money should be spent on making FOSS better rather than making private companies richer thereby helping them kill FOSS.
The last 10 percent are a very hard thing to do - a lot of effort goes into these frigging details (one of them is internationalization, like a proper spell checker for libre office that works for many languages and not just English)
Came here to say this. That last 10% of the polish easily takes 40+% of the effort (or more based on the pareto principle).
GUI work in general is a lot of work.
Why would you want to recreate Windows UI? It's an evolutionary dead-end and everybody is moving off of it. KDE is a dead technology as well, they're still stuck in the WIMP metaphor, whereas Gnome are pushing the envelope with innovative tablet-friendly user-centric story-based UIs that are miles ahead of anything Apple can dream to achieve. We can't move forward by being stuck in the past and an effort to hobble the F/OSS desktop by EU governments should be resisted at all costs.
KDE is a dead technology?
Technically, it uses the latest Qt and CPP world have to offer + QML Declarative UI's and is among the last projects pushing for a community based mobile operating system, designing a touch-based mobile-specific UI.
The only difference is that the desktop metaphor is not being manipulated to become a one-fits-for-all solution.
Instead, a whole new interface is being created for mobiles.
I don't see how coherence is a problem. Windows the platform is pretty incoherent.
The article is light on details, but I think the issue is really in interoperability. When everyone around you runs Windows, you either have to be very good at interoperating with key Windows apps (all parts of Outlook, Word, Excel, Visio, MS Project, etc), or to run Windows, too.
Part of the problem is that many of these things operate on closed data formats, and MS has little incentive to help with interoperbilty. Conversion tools can exist but they are often cumbersome, imperfect, and thus painful for regular use.
This is how an ecosystem lock-in works. BTW it's not specific to Windows; every platform makes it simpler to use this platform's tools, and, because of that, any out-of-platform tools become harder to use in comparison.
True. It's something I didn't write in the original comment, but Office suite is another big reason for Windows being this prevalent. For most non-technical users, MS Office is the only program they know. for such non-tech users, Google docs is the lightweight office you get if you don't have the privilege of running fat Office on your desktop..
Well, technically Linux ecosystem has Open&Libre office. Many times when they show up in news feed, I download and try to use it. But, one thing thing I can always feel - awfully slow/laggy and looks like they are stuck in a decade ago and refuse to change. This makes it uncool for anybody to switch to it from MS office. Perhaps it is great for those obsessive tech users who tinker the heck out of it, but out of the box, it IS inferior in most ways. If anything incentives like these huge government implementations have the potential to start and fund a project that will develop an open & usable alternative to MS Office. One can dream.
> we need a company other than RedHAt
Why not government? FOSS is for free public use like any other public infrastructure. It benefits all so it seems proper for government to fund it.
EDIT: If anyone wants to downvote because they disagree (not that I think that's proper use of the downvote), please at least state your opinion. I want to know.
I totally agree.
I have come to realize that capitalism is a narrow-minded approach towards progress. It does not work in the long run.
Capitalism and democracy are incompatible.
Well, I disagree that capitalism doesn't work, but I can agree that it doesn't always work. Like in many other things, balance is key.
> an opinionated but coherent desktop
Isn't this Ubuntu?
I experimented with Ubuntu by giving a laptop running plain Ubuntu to my retired parents. They are new to computers and their needs are simple.
* Video calls - Skype or FaceTime
* watch youtube.
* Transfer pictures from digital camera to laptop and view them (many times ;-) )
* local newspaper's website
* play their local music library
* Not confuse them.
But, Ubuntu failed them miserably. They had a hard time using because, their words - "what works one time, doesn't the next time."
When my dad closes the lid, he expects the laptop to go to sleep mode, and again open the lid to wake it, just doesn't work all time, give some kind of messages only a tech person would understand.
Their observation - "text is not sharp". I checked it on their laptop display, Windows is rather sharp and usable while Ubuntu's default font rendering is too soft and they need to squint a bit to read.
internet connectivity - There are so many issues with the default wifi setup on their laptop. Dad gets frustrated if it shows him a dialog box with information he does not understand. So, often times the laptop lies there gathering dust until I or a friend goes to their house to "fix it" - often times just reboot the damn thing.
After this struggle, they bought a Windows laptop.
I feel like there are two models for desktop computing—one that is very hard to accommodate for a traditional enterprise, and another where Linux is very well suited.
If one is looking for a managed solution with heavyweight apps including an office suite, email / calendar, etc., then anything other than Windows is going to be unsatisfactory to people who expect Windows. I even include macOS with Linux in that assessment. Let's call this model of centralized control of fat clients the culmination of the 1990s enterprise desktop.
On the other hand, if a company has gone down the road of Google / web apps for everything, then Linux starts to be a very effective bootloader for a web browser.
Personally, I like the fat-client model of the 1990s, and feel that Linux best satisfies my desires. But it's hard to deny the comprehensiveness of Windows & Active Directory for those enterprises that also want to stick to this model.
The web-only model feels a little too lean to me. In any enterprise, there are bound to be instances where web apps aren't going to cut it.
This is where I have hopes for Chrome OS once it begins to support native Linux apps. At that point, it's not a thin client, and it's not a fat client. I wonder then if we may see some of these enterprises walk back from Windows.
I think we are a long way from anyone walking back from Windows. It will be an uphill battle for Google due to the lack of privacy awareness (I'm not saying Microsoft is better or worse, just that Google is not looked upon kindly in the privacy regard)
Perhaps. But I also think that, of the people who are sticking to the heavyweight fat-client model of computing, some are doing so for technical reasons and others are doing so for principled reasons. If Google's softening on the "everything must be a web app" removes the technical barriers, I could imagine an expansion of Chrome OS.
As a paying FastMail customer, I'm right there with you on Google's privacy posture. But given the number of corporate G-Suite customers, I don't think principled rejection of Google over privacy concerns is the major impediment to Chrome OS adoption.
I was under the impression that most of the administrative work was now done in a browser anyways. I wonder what exactly they get by switching to windows, better fleet management tools maybe?
MS is extremely flexible with pricing in their EAs. They can make any arbitrary product pretty much free by gladly readjusting prices, discounts and benefits in other parts of the agreement.
It wouldn't surprise me if they made the switch appealing by this sort of shenanigan, but client CALs are pretty much free already at volume. It's the server licenses where they kill you ... but with Azure flexibility as an option, not to mention O365, it's super-easy to see how MS makes lock-in look very, very appealing.
source: use to own the MS EA at an F200 company that tried to extract Office in favor of Google Drive + LibreOffice, and Windows Server + SQL Server for CentOS + PostgreSQL.
I think the assumption is incorrect. The Microsoft Office suite is used a lot, IIUC
Concur. The office suite and surrounding ecosystem (Exchange, Sharepoint) are king in enterprise environments. The latest versions run poorly or not at all on anything other than Windows.
But the linux alternative for those tools are exceptionally decent and free.
I'm not a heavy MS Office user. But at my current job I get to use Office 2016, while at my previous job I had to use LibreOffice. While LibreOffice is usable its just nowhere near the quality of MS Office. I take this as an example, because I think this is the tool most people will be talking about :).
As a programmer my company shelves out a few hundred dollar per year so I can use a high quality IDE. In the end this saves time and money.
If I would be a business analyst or PO I would expect the company I work for to shelve out a few hundred dollar per year for the apps I need in my day to day work. In that case it would be MS Office (and Windows).
(MS Windows Enterprise edition is only $90 per year, Office $360, that's a lot less expensive than my developer tools).
Good tools save money, and currently I can't think of anything that surpasses MS Excel, MS Word, or MS PowerPoint
Open/LibreOffice compared to MS Office is a joke in a corporate environment where Excel and Powerpoint are kings.
There is also not a single email client that integrates resource management as well as outlook good luck linking Thunderbird to your conferencing and meeting room management systems.
What’s a better Linux alternative than Outlook? I’ve tried Thunderbird and I’m not impressed.
I’ve noticed that a lot of Linux users equate “free” (as in price) with “better”, and this isn’t necessarily so.
Maybe but it's like showing up with a shiny new street Mercedes at a Formula 1 race. You'll just end up being the safety car.
Don't take each solution in a vacuum. Consider that you have to use one and put them head to head. Unless you're a very flexible company or with some very very niche requirements the MS solution will win every time.
People usually understand "Linux (open source) is no good" instead of the real conclusion: today it's not the best for everything so why assume it should be everywhere?
I think this is a common situation: a worker is sent a non-trivial Word document, possibly from someone external, that will only be properly rendered in Word itself.
Even that moves to the cloud though. I have not used the office 365 tools at all but from my understanding they should have feature parity at some point.
But then I suppose that editing a hundred page call to bids document in a browser tab does not really work well.
I'm under the impression that its more to do with external hardware compatibility like printer drivers (do industrial printers have linux drivers?), license plate scanners, etc.
I don’t know about license plate scanners, but usually the more industrial the printer, the higher likelihood that Linux will do just fine.
I think the main reasons are that big industrial printers are likely to support standards like IPP and Postscript. Also, Apple uses CUPS just the same as Linux. Some cheap home printers require proprietary binary drivers, but many printers are supported by CUPS with some minimal config definitions.
Linux Foundation needs to run a hardware support testing program for desktop linux to be successful. In a situation where hardware manufacturers don't support linux and there's no regression testing for community developed drivers, it will never be successful.
In my opinion, Munich's Linux project could have been saved if they invested a little in the UI part. It was horrible out of date, a pain to use and looked like something out of a 90's geocities web page.
The work flows were slow and had too many unnecessary clicks.
If Linux (or somebody) would focus on the UI and stability in a the non-perfect world of the dumb tweaks I do to change the gem color on my flux capacitors, I think it would be viable, but right now, Linux feels like owning a fish tank, and I can't do business on a fish tank, even though I really want to.
They will heavily regret this decision.
If MS was offering them "cheap N years contract," than it is already decided that they will jack up prices the second their contract expires.
> They will heavily regret this decision.
Unlikely. Yes, this is a political decision as most actual users of the systems were ok with the linux client. However, the article is a bit misleading. They are bringing the systems used by a subdepartment of the states department of finance in line with the stack used within the remainder of the department (about 1/3 were linux clients, the state already owns over 50.000 other windows clients (lower estimate, most likely more if you include organization which manage their own infrastructure (like universities)), they know what to expect from microsoft). While the systems worked it always required special attention when interoperating with other states (for example: certain informationrequests by other states are standardized between all states (based on a shared windows solution). Lower saxonian employees could not interact with it directly and had to rely on their inhouse solution instead of reusing the existing and proven solution used by all other states). There are notable costs associated to beeing a special flower while maintining conformance to financial regulations.
Yes, I believe that biting the bullet and migrating more departments to open systems would be the better longterm solution (which would also force vendors to port their solutions). But there will be several elections before that pays off and would require a significant investment before the benefits show (and possible coordination between the states) and in the midterm moving to Windows might be the better option. Moving to linux doesn't win an election, but failing to move to linux might loose you one.
However, even after the move about 1/4 to 1/3 of the computers/servers owned by the state will run linux/bsd (iirc).
> This time it's all about the roving staff, who aren't using OpenSUSE and didn't use Solaris before it. They've always been Windows users and it seems that officials feel it might be time for a bit of consistency.
So my takeaway is that someone at the top was pissed off that he had to use an e-mail client other than Outlook and made a multi-million dollar infrastructure decision based on that.
> So my takeaway is that someone at the top was pissed off that he had to use an e-mail client other than Outlook and made a multi-million dollar infrastructure decision based on that.
Nah, decisions like this are usually greased much better than just by the good feeling of using Outlook again.
I've worked on a number of large rollout projects where multiple vendors have been considered.
When you tell a CEO/CFO/CIO that you can bundle OS, updates, productivity tools, communication/colab tools all in to one monthly bill they almost always take the MS route. The predictability of the financing is the key thing helping MS kill it in enterprise.
It also helps that 90% of their staff have never known anything else and will need very little training.
> very little training
That's the worst assuption ever. Not because everyone uses tool X, everyone understands how to make it work properly. Did you ever try to move a picture from a document you got from M. Johnson? You'll get a headache because moving that one picture will fuck up your title formatting in page 13 and the table in page 2.
Everyone using MS Office I met could use a good week-long training session about do's and don't's, how to write proper documents, presentations that don't suck, etc.
Well, the users usually think they don't need training, and isn't that all that counts?
And to be honest they're not wrong (in those aspects)