W. McDonald Buck, retired CTO of World Bank, finishes his four part essay on the subject of why Corporate Desktop Linux is an unrealistic goal in the short term for Linux advocates. “The hard truth is that the benefits that are most important to individual technical people are simply not important to those lacking technical skills. When you couple this with the relatively meager hard dollar cost savings, the prospect of some extra costs of migration, and the large risks of such a move, is it any wonder few corporate customers are making the transition?“
Part IV & Final Corporate Desktop Linux – The Hard Truth
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I hate to use the word, but that’s what this series of articles was.
I respectfully disagree. The guy is simply objective, based on the way his IT department would take decisions on his line of work. He speaks out of experience, I don’t think that he cares if Linux is the best of not. He just writes what he thinks would happen and I really liked the series. The fact that he was not positive on linux doesn’t mean that the guy wasn’t objective.
Yes we’ve learned both here and on Slashdork that anything that critisize Linux as an option for any certain task is FUD….
If you actually bothered to read the article and figure for yourself that Windows isn’t such a huge investment, you’d realise it’s not FUD.
Costs however, are hardly related simply to the OS, it’s so many other things, such as what you run on top, like MSSQL server is very expensive compared to some free options… but Windows can hardly be called expensive.
based on the way his IT department would take decisions on his line of work
But that’s the problem. He approaches it the windows way. His cost savings are based on the idea of getting new Linux PC’s the way one would buy new windows pc’s. Part 1 has him surfing Dell’s site looking for those boxes, and his mistakes continue from there.
He’s right about one thing, no one pays $299 for windows. They arrange corporate licensing and have custom images, including a lot of pay software. They pay for windows, office, mail client, etc.
If (big if) they are not tied into proprietary systems that are not cross platform, a company could just as easily create a lightweight image using something like debian or slack. Picking lightweight free components, they could use current office hardware(let’s say it’s around a 1.5ghz average) for up to 5-7 years into the future.
I love when these guys include training costs. How many of these people were ever trained on windows? They just figured it out (and some never even did that).
Another possibility this guy didn’t face. Take a big freaking box that has raid drives and keep it updated. That could serve desktops and office apps to 30 or so thin clients (want proof, go to a college CS lab). How many years could they get out of current hardware as just thin clients. The only box that would need much administration or upgrading is the central box.
I’m not saying the information I have above is a “study” or proves anything. But I think those are holes in this guys case.
You can’t think like a windows user and calculate linux costs anymore than you can think like an automatic driver and drive a stick…
-bogey
I have been involved in two Internal Audits of IT ppending in two different companies. In both the price of Windows XP on our desktop machines was not even a factor, it was the price of the application, the servers, the support contracts.
Side note: In both companies we have had a hard time finding Linux/Unix support engineers. Most were either total hacks or very expensive.
> He approaches it the windows way.
Sorry pal, but MOST companies that’s how they are going to approach it! The article IS on par of what most companies will think, therefore this article is very valuable.
I’ve been using Linux and Windows a long time, sold about $70,000,000 worth of it over the last 10 years all told, and I’m here to tell you, this article is FUD.
If you buy machines with pre-installed Linux, you’re 99% of the way to the savings, right there. Sure you need someone to support you, etc. but all the driver problems etc. are dealt with that way.
Get past that and throw a few Citrix apps on the servers you already have, and voila, you have a working Linux network. Re-training takes hours using KDE (most people can use it straight away if they’re used to Windows).
Get some headsets, throw in an Asterisk system and have everyone use the phone system through their computer, set up HylaFax on the same system taking and receiving faxes. Set up something like OpenGroupware on an old Pentium III. Take a look at the bottom line and gawk at what you’ve accomplished for the dollar.
Usually there’s one or two apps you just can’t migrate off Windows right away. Thus, Citrix (or just RDP if you won’t have any slowish links to the server) takes care of that.
It isn’t that he approaches it the ‘windows way’, it is that he approaches it the CTO/CIO way for a huge corporation.
They tend to lease computers, or they buy them, depreciate them over 2 years, and replace them. You don’t make the sale to the CFO saying, but you could use those laptops for an extra 1 year if you ran Linux. The associated costs with migration, re-training, potential lack of support for key applications, etc, are much bigger factors.
Corporations focus on their core business. They cut costs where they can, focusing on things like ROI and IRR. They don’t risk their ability to run their business on the diminishing returns, but focus on the low hanging fruit, in this case OSS applications.
Oh yeah, regarding Dell….a huge number of corporations use them. Look at all the Dell laptops used by business people, check it out at the airport sometime, or wherever.
You are right on the training, companies don’t pay people to learn windows, because they hire people who already have experience using windows. During the influx of computers into the corporate environment, the productivity gains from computers was enough to offset the ramp up as people had to learn. Now, you would take a huge productivity hit and have the training costs to switch everyone midstream.
– Kelson
Sorry pal, but MOST companies that’s how they are going to approach it! The article IS on par of what most companies will think, therefore this article is very valuable.
Thanks for the condesending tone.
So you think he should draw a firm conclusion Linux saves a little if anything and leave it there?
Maybe he could say something like, Linux won’t save you a lot until your IT department learns to leverage a better way of configuring the office. So long term savings can be larger and would require more study, but don’t expect short term results.
Real cost management requires time and objectivity. Ask an accountant who knows nothing about free software.
-bogey
Heres some math for you.
1) People who sell $70 m of windows and linux (not sure how you sell that) are sales guys who can’t spell citrix, let alone suggest you use it.
2) KDE is not like windows. Its similiar but your still going to get a flood of support calls.
Nonetheless, its a very good troll.
How much training would it take to get a user as comfortable on Linux as they are on Windows? Keep in mind when answering how many of these people won’t go into the control panel…
-b
After you pay for a Citrix client license and a Terminal Services seat you have basically paid for XP.
As for running OS apps I’m sure no one disagrees with you there.
I disagree with the author on his closing comment: that the disadvantages of Windows (greater vulnerability to hackers, spyware and viruses) are outweighed by its advantages (greater software availability, greater ease of use). There are two problems with his argument.
One is that where I work, a hospital that is as about as Microsoft-centric as possible, the loss of productivity at the desktop between spyware and pop-up ads is staggering. And at the back end, it’s equally problematic for them, trying to work out a way to keep the stuff out at the gateway. Conversely, in my home we have two Macs (one OS-X and one OS-9) and a linux machine (mine, Gentoo). There are no viruses. There are no pop-up ads. There is no spyware. Period.
Number two has to do with usability. I have seen several people, including my own mother, clicking and typing away at a Linux workstation, without realizing that they weren’t using Windows. And we’re talking Konqueror and OpenOffice. In fact, even on some of my clients’ machines I’ve installed Firefox Thunderbird and OpenOffice as direct replacements for their Microsoft counterparts, at their reqeust! If they can tell the difference, they simply don’t care. And I didn’t provide one iota of training for them. They work in the front office of a dental practice, they dove right in and they haven’t lost any productivity because they’re using non-Microsoft apps. And it’s hard to imagine that productivity dropping off dramatically if they were to use those apps on Linux/KDE.
So I don’t understand how the author can make such blanket statements. My experience suggests that he’s all wet.
1) I didn’t say I was the salesman. I’m a business owner.
2) KDE is similar enough that most people can be comfortable within a couple of hours.
My wife didn’t even know she wasn’t in Windows the other day… except that she couldn’t find Firefox (I didn’t install it, I use Konqueror).
Not a troll, I’m quite serious. I stopped using Windows on my desktop in 1996 in any serious way (I used a Sun Ultra 5 for a while, back then).
>After you pay for a Citrix client license and a Terminal
>Services seat you have basically paid for XP.
Almost true. Many other costs go down. As I said, you could just use RDP/TS if you have the network for it.
The thing is, usually the apps you can’t quickly get rid of on Windows tend to be vertical apps that are only germane to the positions of a few people in the company. In our case, that’s been BusinessVisions.
This guy’s upset because you might not save $ in the first few years doing this (due to retraining, etc). Even if he’s right , How many corporate IT initiatives lose millions or more in the first 5+years, but are still undertaken because of long-term possibilities.
Ask any company that’s ever implemented SAP what their financial plan looked like.
I’d say that “breaking even” the first couple of years and then saving money is a great plan…
-b
Could bother reading any of it, but the reason why Linux on the corporate desktop has not ignited yet is attributed mainly the fact that Linux is not targetting another important market, the home user.
You have to understand that the majority of corporate Windows desktop users are also running Windows at home, which makes for a case of having an agile set of users who are already accustomed to one environment at both home and work.
So far, Linux has not made any significant attempts to capture the desktop in general.
Absolutely. Linux also lets you scale one small app at a time on a forward basis, while usually every few years, the Windows (and Solaris, in my experience) way means a complete forklift upgrade and technology migration each time.
A good, though odd analogy: It’s a lot easier to eat a 5lb steak one bite at a time over a period of years than it is to eat it all in one gulp.
>Could bother reading any of it, but the reason why Linux
>on the corporate desktop has not ignited yet is
>attributed mainly the fact that Linux is not targetting
>another important market, the home user.
Pre-installed Linux could change all that. The biggest barrier to home uptake is that users would have to delete Windows to install Linux and go through all the hassle of finding the drivers, BIOS oddities, etc. that many computers have these days.
There’s at least two distributions that have built systems that are close to what is required to do this properly and make money at it… but neither is being installed by the vendors.
Take a look months back about how many of the vendors are ONLY making money from the Microsoft kickbacks that they get for installing Windows, and it’s fairly clear why this isn’t going anywhere in North America at least.
A good, though odd analogy: It’s a lot easier to eat a 5lb steak one bite at a time over a period of years than it is to eat it all in one gulp.
If I eat the steak in one go I know it’s going to be fresh, who knows what condition those last few bites are going to be like after 5 years
Quick…emergency…we’ve gone 5 minutes without a car analogy.
I’m amused by the responses to this series that think this guy is only full of FUD. Much of the author’s criticisms seem valid to me for people making practical business decisions.
I also think most of the problems he points to are things that are going to be solved one way or another (they are surmountable).
For example, in this latest article his point about users being overwhelmed by the technical choices that Linux provides is true, but misses this fact:
Aunt Tilly will benefit by using a distro that chooses a coherent set of applications and presents a coherent interface. Distros with a strong hand on the rudder are going to simplify the choice problem.
Whether that’s Linspire or Xandros or Fedora or something yet to be, is there really much Uncertainty or Doubt that some commercially-supported newbie-friendly distro will meet that clear market need?
My wife didn’t even know she wasn’t in Windows the other day…
Similar experience here: my girlfriend is a computer noob and she didn’t take long at all to learn to use KDE. Gnome is also very easy to use.
It’s interesting I was watching Jamie’s Dinners last night and the things he had to do to get kids to eat healthier….
First success was to copy MacDonald’s i.e. give away things
Second get the kids involved… they make it themselves…
Third and very important was to hand it over via a peer…
If one said they disliked it others would copy….
Seems we can learn a lot from kids
……..and icon is an icon………….
I thought the article was very good and very far from FUD as other posters complained.
The point the author makes is that linux, in companies with thousands of desktops may not really save any money at all in the end when you factor all costs of migration. And in the small business sector, where it should thrive, it doesnt either, since the cost of the technical expertise required handle the migration, setup and support can be as or more expensive than sticking with Windows.
That doesn’t mean linux sucks, it just means that the marketplace is still not mature enough to offer a a full linux enviroment (applications, expert workforce) that can compete strongly against windows in this particular segment. It will happen eventually though as more money is invested in linux.
Yeah, migration costs are the biggest factor. These companies are already using Windows and have been for years. I’ve been dual booting for 4+ years now and have been unsuccessful in completely migrating over to linux. I have nearly 300mb of email trapped in Outlook that I have been trying to migrate over to linux for a good year and a half. It can be done, but as of last night, I still haven’t been able to do it to my satisfaction. Now running an Exchange server, this might be a little less of a daunting task. But migrating away from Windows is very difficult.
My impression is not that the article is FUD, but that that the author is simply “drinking the cool-aid” of a different flavor than the rabidly pro-Linux camp.
Linux on the desktop is an excellent choice for many reasons in a number of scenarios, poor in others. It’s principle feature, from an organizational stand-point is that it’s different. There are different rules, different liabilities and considerations — it’s just different. For a naive organization, “different” poses problems in that the current staff may not have competencies necessary to change (assess, plan, execute, and support).
While the author would have you believe that it simply doesn’t make sense to switch to Linux on the desktop, nonetheless many organizations have and seem to have very positive results (and others have and had equally dismal results). The author make many dim-witted assumptions about switching to Linux (only a few touched on by posters) that color his conclusion, but none of it is contrived. That is to say: if you follow his approach, which is reasonable from the point of view that you are simply replacing Windows on the desktop and not focusing on leveraging the technology or changing your IT processes, then it makes perfect sense.
Just a simple example: one poster suggested that it’s silly to keep replacing machines to keep up with OS upgrades while another pointed out, correctly, that it’s not relevent because most companies lease their machines for two years and swap them out. What the conversation lacked was the observation that tech refreshes are on a
2 year cycle mostly to maintain compatibility with an ever-changing software platform that clearly is designed to require regular “by the bootstraps” upgrading.
There’s plenty of anecdotal evidence that Linux on the desktop is clearly a viable option for many (our local home improvement warehouse uses it, so does the retail outlets for my cellular provider, as does my employer for a number of research functions). There are obviously application dependencies that are still not resolvable/replacable that would make it unreasonable (and even where that’s not true, many IT groups don’t have sufficient Linux knowledge to know recognize it).
This author’s thesis holds water so long as your organization doesn’t apply due dilligence to the change. Management is not likely to change anything they don’t understand or appreciate without being sold by IT. And, honestly, IT groups are overworked and underappreciated already. This would just be one more thing to add to their plate, and in many cases such a switch might be perceived as an effort to downsize the department (different skill sets required, fewer support personnel required, etc.).
Linux readiness for the desktop, and even the economics of migration, really aren’t relevent to whether it’s actually deployed. The article doesn’t recognize this.
I wonder what the cost would be to just switch to a mac instead? At least their support costs should theoretically go down.