Exactly. On Linux you can automatize system administration very well with FOSS. Yes, you have to use different tools, but, for example, SSH-ing into a box with certificate-based authentication when it is up is trivial. Scripting that way is trivial. Even a doing a full installation (on top of a minimal one) this way is trivial. Doing updates this way is trivial. All with the security mind-set that UNIX was designed with from the start.
We were successfully managing 1,500 Windows and Mac desktops with exactlyone Windows Admin and exactly one Mac Admin, along with a DBA/Report generator person to help admin staff get the reports/data they needed.
We had absolutely zero interest in SSHing into individual desktops to run scripts - managing the desktops with Group Policy was a simple, well-known mechanism. I don't mean this as a disparaging judgement, but your comments are really about today, 15 years after the fact - I'm talking about the state of the systems in 2010, not 2025, and your comments make me think you've never seen how a well-run windows shop is administered.
Now, if you want the same admin interface on Linux as you had on Windows, of course you are paying for that. But you are doing it wrong.
Thats your opinion, and that's fine, but that was what we were familar with.
There is a reason you need a lot fewer sysadmins on UNIX (and Linux) than on Windows and that reason is that once the sysadmin has automatized everything, very little breaks. Quite unlike Windows.
Its hard for me to believe you could manage 1,500 linux desktops in 2010 with fewer than two admins like we had. And honestly, the Windows environment, when properly structured and maintained, is not nearly as problem-prone as most Linux enthusiasts claim.
Of course, the other part of this is that UNIX sysadmins are quite a bit more expensive than Windows ones, because the expectations are a lot higher. But you still come out ahead by a lot.
Again, Hard to imagine saving significant money over the then TWO ADMINS for 1,500 desktops, nearly 3,000 users. If, as a practical matter the two existing admins were replaced with two Unix admins earning 50% more (effectively going from two headcount to three, in budget terms) then the complete, $10 savings versus the per-desk annual license cost of Windows is gone. If the two replacement Unix admins cost 10% more, that would make Linux just as expensive as Windows.
Schools are a funny business, yes they get three months off each year (sort of, summer school is a thing), but imagine, as a thought experiment, converting a school district with 1,500 desktops and 3,000 users (student and administrators) from Windows to Linux. Installing the OS on 1,500 computers, that's doable, no fun, but doable. But, there's a whole lot of automation and tooling bundled into Windows Server and Group Policy, much, but certainly not all, would need to be either purchased, implemented, and trained on between school years, And since school districts rely on boring things like attendance records (for federal funding), similar replacement software would need to be found, installed, trained on before the next school year. Oh, and the students AND faculty would, believe it or not, need some level of training on the new software tools AND operating systems without overly impacting their educational progress as they learn the new environment.
AGAIN, this is 2010, the world was not "browser-based", applications were distict local software packages installed on computers/servers.
The only practical solution would be to develop the new environment/ecosystem (OS, Servers, desktop images, application suites, and administration tools) in parallel, over a school year. That means you've got the Win Admin and Mac admin on staff, plus the two new Unix admins, plus a handful of experts in either Linux applications for schools or programmer analysts designing and implementing Unix-based solutions to all the issues run into during the migration.
That implies a huge (one-time) headcount increase - how many years of $10/desktop/year savings would it take to recoup those costs before we start "rolling in the savings"? And, where would the money come from - would the state give us sufficient money to hire several Unix experts for a year or two to migrate to Linux? Would the taxpayers approve a ten-year bond to increase their property/school taxes to pay for these future savings?
In the world of hand-wavy social media, migrating to Linux from windows is easy, heck everything (today, not 15 years ago) is browser-based, so whats the difference? But the reality is a bit tougher.