Where is Largest Linux Desktop Install?
jackb_guppy asks: "Talking with Managers about Linux Training for staff. One asked a simple question: 'Where is the largest installed base of Linux desktops?' My guess the question was asked prove that there is no market, and I am unable to find an answer. I am guessing the next question will be: 'Largest site using Linux to replace MS desktops?' Anyone have a suggestion?" Just for fun, if any of you have Linux Desktops deployed in your department, can you give us some numbers?
They apparantly saved over £250,000 (~$350k) by switching to Linux and Star Office. This was during the uproar about MS licensing.
--ALex
My Poor Sig.
-- My hovercraft is full of eels.
The school I am network manager for has a deployment of 250 linux based thin clients running citrix desktop software off 6 NT servers.
:).
The linux client software is a heavierly customised RedHat 6.2 with an ext3 root fs (which is great not a single ext3 related fault on any of them and no fsck's) that basically boots an X session with a full screen citrix client on them.
All the full clients that we have (50 or so) have a dual boot option to a full RedHat 7.1 + Ximian 1.4 desktop that is used by some pupils and also the art department for Gimp and Blender work.
While these aren't true linux "desktops" they run run linux and sit on desks
All the back end servers are linux baring the 6 NT desktop servers. All filestore, mail, web, dns, auth, etc is run on linux.
I really can't believe a user experience like this is new to you. The very fact that you're hearing it now will hopefully clue you in to what life is like in the non-optimal real world.
Wide-scale Linux deployments have worked for, primarily, two reasons:
1. The new Linux installations were, in general, not designed to replace a totally-flexible Windows configuration. Generally, these installations are more akin to point-of-sale operations, or for configurations that have a very limited range of use and do not need a huge application base for their users to be productive.
2. Those Linux implementations were well-planned, well-thought-out and well-engineered. Generally a small team went through, selected a list of the appropriate applications, built a custom distribution or post-installation checklist and custom-tailored the OS to their environment. Most large corporations even do this with Windows.
Your experiences with FreeBSD and your non-tech staff are not necessarily representative of everyone else's. Just as you find it difficult to believe the original poster's story of difficulties, I find it difficult to believe a moderate- to large-scale conversion to FreeBSD or Linux as a Windows desktop replacement (for all-purpose tasks) as you're describing would go without a hitch or a noticable loss of productivity.
That doesn't mean I wouldn't consider your comments valuable. Clearly some people are having successes in some configurations and others are having problems in other configurations.
As long as they hold this attitude, it ain't gunna happen. You can use ext3 journalling, you can give them the latest-greatest KDE apps or Mozilla (which scream along in comparison to the second-latest-greatest), you can do all of these things to no avail. They will still find something to whine about.
Now if you keep your eyes open, you'll find one or two important people playing a game or using an app that they treasure. It might even be Solitaire. When you do discover what the sticking point is, install it or a better one on Linux boxes not being used by them, and show the other users how to use it. The answer to Solitaire would be PySol. You may lose a few man-hours to a game, but once your legacy system users are acquainted with it, the transition suddenly becomes easier.
A similar tactic is to install one or two Linux boxes in their area ``for visitors'' or whatever, and put a whole raft of really interesting things on them. Then have people go in every so often and use the interesting things on the Linux boxes only. I don't know your people; it might be bzflag, TuxRacer, Jabber, anything. Who knows what pushes their button? Try them all.
Got time? Spend some of it coding or testing