German state moving 30,000 PCs to Linux and LibreOffice - The Document Foundation Blog
German state moving 30,000 PCs to Linux and LibreOffice - The Document Foundation Blog
Following a successful pilot project, the northern German federal state of Schleswig-Holstein has decided to move from Microsoft Windows and Microsoft Office to Linux and LibreOffice (and other free and open source software) on the 30,000 PCs used in the local government. As reported on the homepage...
You're viewing a single thread.
The last time I tried it, which is now a few years ago, LibreOffice Calc was substantially slower than Excell for larger spreadsheets. Like a difference between night and day, it was no acceptable substitute if productivity was a concern, which it usually is.
Imo a big swoop change like this, which is done for ideological reasons, but without practical considerations, is doomed to fail and leave a lasting bad impression in peoples' minds. Imo it would have been far better to only drop windows 10/11 for a familiar looking Linux distro, while continuing to use Microsoft Office.
If an Excel sheet is that big, it should be replaced with a proper database, which most likely would run on Linux. I think you're right, though, about the lack of planning around the practicalities.
Using the right tool in an office setting isn't something that's typically done. Unless that tool is a spreadsheet of course. A spreadsheet is always the right tool.
There often aren't spare funds to develop something proper.
What is the best sheet to Db Tool out there? Surely this is a normalized path by now, right?
you can probably just use python (pandas or something) or an equivalent in other languages
In structural engineering (bridge design etc), we use quite complicated spreadsheets for calculations; a database wouldn't be the right tool for that job. We use excel because everyone knows how to use it and it's easy to print to PDF and see the inputs and outputs and any graphical summaries you have added. Using a spreadsheet makes it easy to check and easy to adapt/change when you want to do a slightly different calculation next time.
I've tried building spreadsheets of similar complexity in libreoffice and it's true they are very slow in comparison and more prone to crashing.
Libreoffice works well for some tasks and I enjoy using it at home but honestly if I tried to use it at work it would cut my productivity significantly. I'm probably using it more intensively than most people though.
I agree that spreadsheet use in engineering is one of the most complicated use cases, but I submit for your consideration another very complicated use case: laboratory software ( https://en.m.wikipedia.org/wiki/Laboratory_information_management_system ) LIMS do what Excel can but with the added benefits of being more controlled, secure, user friendly and faster because they're built upon the back of a modern database. In my experience with engineer built worksheets, the engineer that built them is typically the only one who knows how to use them. This is job security for that engineer, but isn't scaleable for others' use. In the lab software, a scientist builds the methods, and lab technicians use those methods over and over again daily. Each step of each use of the method is recorded with the inputs, the results, who performed it and exactly when. The workflows are built-in and the calculations are comparable to those used in engineering.
Interesting, I think it's different for structural engineering because you're doing calculations in accordance with a code of practice and the spreadsheet needs to be adapted to tweak the inputs and outputs of a standard formula and apply it slightly differently for different bridges / structural arrangements. I've written loads of spreadsheets that have been used and adapted by other people in my company, I honestly don't think they are that difficult to understand (or people wouldn't have been able to build on them and adapt them).
I can see that lab software is quite different, especially if you have very well defined procedures and you are repeating exactly the same test again and again with the same inputs and outputs.
Ah, yes. Easily adoptable by coworkers + low repeatability = no need to change. Stick with spreadsheets.
Software freedom is about being in control of your computing. We can't verify what proprietary software from a foreign company is doing on a government computer.
Public money public code is about citizens getting back the code we paid for. When a proprietary company improves software to get paid then they keep that advantage to themselves. LibreOffice is a collaborative project, everyone gains from it being improved by our money.
These aught to be valid concerns as much as productivity: to the degree it affects people. It cannot be dissmssed as being idelogical.
Those are ideological reasons though and me calling them idealogical does not mean that I dismiss them as valid reasons. Idealogy in itself is not a bad thing and it should certainly have a part in decision making.
Where we differ in opinion is in which should take priority: I'm of the opinion that practicality should trump ideology (in this case), while you find the idealogical reasons more important.
Are you working on ods formats? Anyways... We need more people using LO so we'll have more developers and rhen comes a point in which LO surpasses everything else. Meanwhile, if we stick to proprietary software,we would be stuck in a vicious loop. We need to break trough and sometimes inthat transition there are some concessions to make. If we manage to make it aa a big collective of people,the transition process will be shorter.
I'm not using it anymore, I just tested it to see if I could propose it as a substitute. In my testing I tried both open and ms formats: I started with old excel files which didn't work well, so then I tried open format files that were build up from a clean slate state, with the data imported from CSV files. After that didn't perform satisfactory either, I turned to the internet. After searching for the major issue that I encountered (slow in a large sheet), I came to the conclusion that calc could not be a full substitute for excell, so I never proposed it and we're still using ms office to this day.
I'm just going to copypaste some other people's thoughts with which I agree, saving me a bit of time:
*"If you work at a large company for a while you'll encounter a class of user that Calc doesn't really address. They're like super-specialists. They often have a deep knowledge of Excel, but are otherwise completely computer illiterate. They also work with large datasets and specific models. Calc isn't a replacement for them. Not just on a feature level, but on an accessibility level.
Look for Excel resources. Classes, books, articles, howtos, everywhere. Do the same for Calc and you'll struggle a lot more. There is stuff there, but it just isn't nearly as professional and rich. There is no great way to transition Excel users to Calc users and have them still be as productive.
In the Linux world, when we get those style of work-loads we generally put aside Calc / Excel as a tool and begin looking at programming languages (e.g., Python, Matlab). I feel like this somewhat handicaps our ability to reach those users.
for basic use though, it's perfectly acceptable. I just wouldn't consider it a poweruser tool, and those power users are what make Office a multibillion dollar product for MS."*
*"Sadly, it's just not there in book.
The only time I try to use LOCALC is when I have a few hundreds/thousands of rows of formatted values to sort into a simple graph and performance is just abysmal.
I just tried again earlier this day and though most daily features are there for your regular user, all my "casual" uses of it ended up underlining the severe performance problems.
Maybe my uses are far more corner case than I believe..."*
https://old.reddit.com/r/linux/comments/9yjwyf/is_libreoffice_calc_truly_a_worthy_replacement/
Wow. i dont know. Ive never used calc nor excel for things that big. For that Ive used python or php/mysql. I use spreadsheets a lot, even complex ones, and calc always work for me. Anyways, you are right in regarda to the lack of support/books. Of course. Same happens with a lot of things. They are not mainstream (yet) because they are projects that havent had the money or power corporations like microsoft has had for lobbies and marketing. So we can struggle a little on the transition (some very brave pioneers have already paved the way for us, so its not that hard anymore) and hope we are contributing to a libre/free future in which digital technology helps build a better and fairer society for all. And then, they'll become mainstream and we'll have more books, support and communities than we ever dreamed of before.
I think we should aim for what is ideal and then take into consideration the constraints of what we can do. If it's not plausible enough to go for what is ideal then aim to make that more likely while doing whatever is the next best thing. We risk being stuck on a peak of possible good if we refuse to go down to eventually go up higher.
I only use LibreOffice but don't need it much. I can't comment on how practical it is, or isn't, for use in a government. If there's another free software option then we aught to consider that. Else spend money to make it good enough for frequent and important use cases.
I think reducing their reasons to ideological is not fair. They stand to save a lot of money, reduce the risk of leaking data (to MS or hackers), and will have the ability to fork/add their own features.
While I am not familiar enough with Calc or Excel to comment on the speed, I imagine having an entire government using it could get the ball rolling on optimizations.
They stand to save a lot of money
Do they? They'll now have to start training people to use Linux.
These people aren't going to be enthusiasts like us. They need to be shown where everything is.
The money that will be saved is peanuts compared to the cost of the workers. Loss of productivity through the implementation of bad tools can be very costly. The various Microsoft Office programs also offer the possibility to add bespoke features. Microsoft Office does not leak data unless you chose to let it do so, at least in the eu.
Optimizations that might happen once a program with unacceptable performance is in a production environment, are generally optimizations that never happen. I've never seen a program make such a turnaround, it's wishful thinking without a basis in reality.
This thing really is set up for failure. I'm not against organisations moving away from products from large monopolistic companies, rather the opposite, I'm very much in favor. But if the move is done in such a way that it's bound to fail and then cement itself into people's mind as a bad thing, then it has accomplished the opposite of what it has set out to do. Right now Linux is ready for widespread adoption in environments where productivity matters, but in my experience libre office is not.
the possibility of bespoke features Such a shame you can't do this with open source software.
Every time I see someone say 'I'm actually really a fan of open source' it reads like 'I'm not racist but'.
It's a question of both expenses and dependency on a monopolist.
There simply won't ever be an opportunity to move from MS solutions to FOSS solutions which won't have these problems.
Being dependent is possibly more expensive in the long term too.
If you throw even half the money that would go to ms license for the foss community instead you can get some pretty huge improvements for that foss program. Blender for example, got actually nice looking and seriously good program while being foss because they got decent funding
Well, people blamed old (archaic, what it had when it was an Amiga program) UI for being hard to use, but the new one is even harder, so dunno.
I touched Blender with the old UI somewhere in late 00s on Windows, managed to sculpt and render a few clumsy objects. I don't remember how long it took, but it feels as if the new one took twice that for the same.
EDIT: On the actual subject - yes, that too. I sometimes think that (moderate) positive inflation is not always better than deflation. It encourages a narrow way of thinking where we always stop at first local optimum. Say, MSO is cheaper right now than LO - then we choose MSO, period. Nobody thinks about finding a bigger optimum, because constant inflation psychologically encourages you to think that way. That's just clumsy philosophy.
I find the new much so much more easy and clear
Lol, has sheets the size where performance matters and talks about productivity. And then chalks LO up to failure because of this. Maybe converting your .xlsx to .ods would save more money then, LO is faster there, because .xlsx & co are inherently broken.
Wow, Schleswig-Holstein meaning well and starting beef by doing it. I'm sure there's no precedent for that.