Slashdot Mirror
Constructing a Windows-Less Office
joewakeup writes "This article at CRN analyses why today is the best time to consider building a pure Linux information system, from servers to... desktop. Among all the arguments, one of the arguments is the low cost of Linux offerings compared to Windows based-solutions. Worth a read."
I work in an electronic music studio. I'd love to use Linux, but the apps just aren't there.
The fact that there's almost no development community addressing this potentially enormous market amazes me to no end.
But, until then, I'll use Windows. Not because it's great, but because it has the apps I need.
As reported in Slashdot this morning, Evolution 1.0 Released and ThinkFree Office an MS 2000/XP Office compatible suite that works in Linux. Combine these with the TransGaming's WineX software, there is no longer any reason to use MS on the desktop.
"I'm The Bounty Bear. I will find him anywhere. I'm searching."
But which software do you run on the server side? With Exchange all the email in a database file and you get single instance storage to save space. The user directory is in another database. Exchange is almost like a slimmed down SQL server. Even Oracle is going to this model for their new email server because it's awesome. Only time we touch our exchange servers is to add users.
Since it's a database you have option called deleted item retention time. When a user deletes an email it stays hidden in the database for whatever amount of time you specify. If you have the storage it could be months or years. If someone deletes an email message restoration takes seconds from the client PC. Exchange 2000 takes it one step further to the mailbox level. If you delete a mailbox by accident or after a termination, you can restore the mailbox with all contents.
Backup is easy. And Veritas even has an option called brick level back up which backs up the mailboxes individually. A company called Commvault takes this to a new level and can back up single messages. If you CEO deletes an email you just restore the message. If you delete a mailbox and someone needs it you just restore the single mailbox.
Is there anything that runs on Unix or Linux that supports these features? On an enterprise level they are a requirement.
It seems like every year I get infected with the pro-linux bias of slashdot and rip Windows off my machine.
I ripped Windows off at about the same last year and installed Linux. I wasn't impressed. The desktop managers seemed slow (I was running a P3-800) and the web-browser sucked and generally, the applications weren't as good as their Windows counterparts. Not to mention that I managed to crash the system and have ext2 throw away some files.
So, this weekend I tried it again. I ripped off Windows 2000 and installed RedHat 7.2. In one year, Linux (and Gnome / KDE) has improved ten-fold. The KDE browser rocks, KMail is very good and the ext3fs filesystem is much better. However, it still took me hours to get ADSL PPPOE and a VPN client up and running and the soundcard (VIA 8233) and tv-card (Brooktree) still don't work. Apparently, the concept of writing a device driver without patching the kernel is still impossible even though Windows/Mac have been doing it for many years. And the system (now an Tbird-1.33) is still slower than Windows 2K (ex., the mouse gets jerky when my apps thrash the disk).
I'm a developer, so I'm thinking of writing support for some of these things (such as an easy VPN installer). Or, maybe a universal driver installer that would automagically patch the kernel and say 'You must reboot now', ala Windows. But the thought of having to support different distributions and versions makes me cringe.
Alot of the problems in Windows can be attributed to Microsoft trying to be backwards-compatible. But with Linux, the kernel and major libraries (ie. glibc) are always changing underneath your feet. This is a major design flaw that I not sure can ever be rectified.
Jason.
It's a good open source project. The initial version doesn't have to support animation, but design in the hooks, and it will probably be added by others. Perl code to read and write Flash exists, so there's something to look at. A good student programming project.
Things "get into" the office environment when they make business sense to do so. Which happens when the benefits exceed the costs, the reward exceeds the risk, and when these are exceeded by an amount greater than the next best alternative.
In the case of office platforms, the big "corporate IT" issue re this analysis in representing the complete true costs - Total Cost of Ownership - which includes the relative expense of good Unix sysadmins or the cost of retraining Win admins (clue injection), the cost of managing the environments, the cost of supporting moronic end users, the costs of reduced application availability (sure you can have a nice GUI, but where's the Linux industrial-strength Accounts Payable system?), or of building interfaces to whatever the rest of the world uses (eg., the cost of reverse engineering .doc format for word processing). The actual cost of the OS (free beer) is almost irrelevant.
On the risk side, corporate IT departments value stability of the infrastructure above all. So, the corporate IT folks are herd-following conformists. No one will move to Linux office until everyone else does. And there will have to be a huge TCO advantage before that inertia gets overcome.
It's actually a rational position, but not very cool or fun. Sticking with the herd, and moving en masse with the herd has advanatges. The herd is big enough that it gets what it wants: robust techinical support, business applications developed for the platform of their choice, peer groups and conferences in Boca Raton, whatever.
Of course, you lose out on the advanatges of doing something different/better than competitors. It all depends on what you value more.
(PHB off)
Just kidding of course. This was posted from a Linux system hiding in a 50,000 person company.
Since these boxes are going to be on a network anyhow, why not simply use underpowered machines as X terminals. Any machine with 16M of ram has enough oomph to be an X terminal, and one commodity server (with enough memory) can easily support hundreds of users.
As far as I am concerned the only real reason to take a look at Linux on the desktop is that it finally allows the systems administrators to move to a useable thin client arrangement. Imagine the joys of one box to administer and nothing but disposable machines on your users' dekstops. Linux has finally gotten to the point where it has enough applications to allow you to shift to this sort of a setup today. The fact that this sort of an arrangement will probably save you money on both software and hardware costs is nothing more than icing on the cake. The real potential for savings is in administration costs. All of a sudden you can get rid of all of your desktop support personnel and replace them with one Linux admin, and a monkey whose sole job would be to replace failed thin clients.