Part I: Corporate Desktop Linux - The Hard Truth (OSDir.com)

Why does the open source community have to demonstrate *anything*? If a company wants to investigate ways of controlling costs or determine ways of gaining advantage over their competition by using different software and methodologies, then let them do it! The "Open Source Community" doesn't live or die based on what corporations and other businesses choose to use as their desktop OS and applications. When they see that they can save money and/or get more out of their systems, well, then they'll switch. Until then, why care? Whether it's laying off 10,000 workers to improve the stock price, moving plants to China to break unions, or investing in new technology that doubles output, Business has never seemed to have a problem choosing the path that maximizes profit.

There's at least one important notion here: pro-Linux BS is BS, and we should (and can) back ourselves up with only real claims. People have claimed Linux to be easy to use (true), to come with "everything you need" (true in some cases, false in others), to be ultrasecure (not true by default, but potentially possible), and to be nebulously "ready for the desktop."

Linux is easy to use and usually comes with enough for a business and the average user at least. For gamers, people grasp at straws: Doom3, Quake3, ok. Older games like Quake1,2 and Duke Nukem, original Doom, etc, sure. New hits like WoW or FF11? No, sorry. Often people are even attached to programs that they have to be eased away from.

"Ultrasecure" claims are made all the time. Linux is not immune to viruses, it's just harder for a normal user to infect the machine. Viruses that take over local root exploits can and will spread when Linux becomes a target for many viruses. I should probably steal the Bliss source code (a virus-on-linux POC) and one of the recent kernel root elevation POCs and hack up a "Linux Virus Proof-of-Concept" from the two. Wouldn't be too hard: int my_attack_routine() { get_me_root(); bliss_me_around(); } Of course, with a little work a lot of this can be helped.

"Ready for the desktop" claims are also often made, but not substanciatable. New hardware pulls ahead of Linux. I have a USB WIFI (netgear WG111) and a PCMCIA WIFI (adm8211 based DLink DWL-650), neither of which work. Newer graphics hardware from ATi and nVidia has no OSS drivers, and the closed ones impede security work. Programs can't be easily downloaded and installed right to the system by a user yet either, something a lot of people do.

And of course, there's some ways to go to get real security, which ties into "Ready for the Desktop." We DO need clamav GUIs that sit in the Gnome system tray and test on demand, as well as host-based firewalls hooking into connect() for the user. Eventually there's going to be spyware and viruses and worms. Heafted-up security is going to be able to bounce the portion of the problem not propogating through e-mail, DCC, and programs on web sites. Digital signing of binaries will help squelch virus propagation.

To substantiate many of the common pro-Linux claims, a lot of work has to go in to make Linux not only a better designed OS in general in relation to current solutions, but a much better OS in a hostile environment. This ties in with TCO, because once Linux hits mainstream, viruses and worms will start popping up, which will severely increase the TCO just as they do on Windows. Needing an expert to chose your hardware for you, or needing to hire driver writers will also increase the TCO.

Mr. Buck may or may not have a point with the TCO article here; but it is true that there's a lot of mystified misconceptions about Linux going around. In both directions.

Look, if you were a slave and yearning to be free, and people were telling you how to get free, would you listen to other people telling you that you needed to wait to be free until the people telling you how to gain your freedom demonstrated that freedom presents a better cost/benefit case than slavery?

Not everything is money. I know I am much happier as a Free Software user than I was before I made the switch. (And here is a little secret, it is costing me less as well.)

Has anyone undertaken to do a legal cost benefit analysis for home and small businesses? Thats is, survey typical windows and linux desktops. Add up the costs if all the software on those desktops was legally purchased and compare. Also, if not all the software on said desktops is legal and provably so, what is the dollar figure at risk for the users. (I am not asking for the likelyhood of having to pay, I am asking how much is at risk in a worse case situation for an average desktop.)

all the best,

drew

I always wonder.

Me, I am a corporation. I have a desktop. I use linux because;

It has a better cost/benefit ratio than *anything* else.

It works when others don't.

I can do things with it others can't do.

The recent IBM paper "Linux Client Migration Cookbook:
A Practical Planning and Implementation Guide for
Migrating to Desktop Linux"
http://www.redbooks.ibm.com/abstracts/sg246380.html
is a very nice read.
TCO is not an issue. Free is free. Paid support is paid support.
On my level, supporting client networks of up to about 100 users,
(but mostly more like 10) all of the usual stuff about management doesn't
even apply. A copy of MS Office costs a lot of money, 10 copies cost
10 times that much. A copy of Open Office costs a whole heck of a
lot less, even with a 1 or 2K donation.

There is no contest.

Yeah, maybe migrating the entire world-wide automotive industry
off of Windows might be a daunting task, esp considering the
deep investment they have in IT skills and knowlege on that
platform. So what? Sooner they do it, the sooner everyone benefits.

That aside, Migrating the local library is a no brainer.

TCO? there is no argument.

As to buying stacks of machines without windows preloaded,
Try ignoring that irritating guy on TV, and go visit your local
whitebox shops and find the local guy who is going to actually
offer you service, and buy from him/her. That's where you can
get machines without windows preloaded, and a smaller MS tax.

More work initially, sure. But worthwhile work.

On another side note, I'm finding that "fork-lift" migrations
are easier than incremental. Folks whine and carp about things
for a few weeks, then they are just working again. Every once
in a great while, a spreadsheet misbehaves, and the end-user
has to ask the author to save it as an older version and resend it.
That coupled with no activeX web-desktop contols, kinda puts
a few so-called web-services out of reach. But in the end,
if that's as clueful as the service provider is about "web" technology,
then no doubt,there is someone else out there providing a similar
service that will get the job done that doesn't require Internet
Explorer 5.5 or better (on Windows, macs need not apply). Choose
your vendors, and get back to work.

With thin clients accessing a Linux instance (pick an underlying infrastructure... RLX, VMware, etc.), a company's divisional or departmental employees can have all the functionality they need to run the business without having to waste money on a low-end server class machine running a powerful virus/spyware/malware magnet sitting on each and every desk.

Why not take some of that money that's wasted on Microsoft CAL's, desktop hardware, extra desktop support staff and/or contractors, virus software, vendor lock-in (which requires you to buy stuff you never counted on having to buy in order to get product A to work with product B), etc., spend part of it on the underlying infrastructure, and allocate the rest on thin client hardware? I realize it won't all even out immediately, and this would need to be a multi-year project to allow currently leased machines to finish out their life-cycle, but wouldn't the long-term benefits warrant such a project?